132
Analytics © 2020 General Electric Company

Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

  • Upload
    others

  • View
    7

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Analytics

© 2020 General Electric Company

Page 2: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Contents

Chapter 1: Overview 1About Application Analytics 2

Prerequisites for Getting Started with Application Analytics 4

Chapter 2: Workflow 6Analytics User Typical Workflow 7

Chapter 3: Alerts 8Generate Multiple Alerts 9

Chapter 4: Asset Filters and Configuration 10Analytics Configuration 11

Manage Asset Filters 13

Chapter 5: Analytic Templates 17Manage Analytic Templates 18

Configuration 26

Chapter 6: Analytic Orchestration 42Manage Analytic Orchestrations 43

Manage Analytic Data Flows 45

Chapter 7: Analytic Deployments 49Manage Analytic Deployments 50

Select Asset Applicability for Analytics 53

Map Analytics Inputs and Outputs to Tags 57

Manage Analytic Iterations within an Orchestration 67

Manage Deployment Schedules 68

Review Deployments 74

ii Analytics

Page 3: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter 8: Analytics Troubleshooting 75Cannot Access Application Analytics Module and Apps 76

Cannot Access Application Analytics Module and/or Features 76

Cannot Access Custom Properties For Alerts in Analytic Output 77

Cannot Deploy an Analytic to Predix Runtime 77

Cannot Access Attributes in the Asset Filter 78

Failed Analytics Deployment 78

Chapter 9: Reference 80Analytics Developer Reference 81

Sample Analytics 98

Understand Analytic Ingestion 105

Chapter 10: Release Notes 114Fourth Quarter of 2019 115

Second Quarter of 2019 116

First Quarter of 2019 116

Fourth Quarter of 2018 118

Second Quarter of 2018 121

First Quarter of 2018 122

iii

Page 4: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Copyright GE Digital© 2020 General Electric Company.

GE, the GE Monogram, and Predix are either registered trademarks or trademarks of All other trademarksare the property of their respective owners.

This document may contain Confidential/Proprietary information of and/or its suppliers or vendors.Distribution or reproduction is prohibited without permission.

THIS DOCUMENT AND ITS CONTENTS ARE PROVIDED "AS IS," WITH NO REPRESENTATION ORWARRANTIES OF ANY KIND, WHETHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TOWARRANTIES OF DESIGN, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. ALL OTHERLIABILITY ARISING FROM RELIANCE UPON ANY INFORMATION CONTAINED HEREIN IS EXPRESSLYDISCLAIMED.

Access to and use of the software described in this document is conditioned on acceptance of the EndUser License Agreement and compliance with its terms.

iv © 2020 General Electric Company

Page 5: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

1OverviewTopics:

• About Application Analytics• Prerequisites for Getting

Started with ApplicationAnalytics

© 2020 General Electric Company 1

Page 6: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About Application AnalyticsIn this section, analytic users such as data scientist and process engineer can find an overview of theApplication Analytics.

The Application Analytics provides a cloud-based distributed computation application suite to provide asingle place to package, configure, validate, deploy, schedule, and monitor analytics. It provides aframework to deploy and run analytics on the respective runtime environments such as Apache Spark,SmartSignal, Predix, and Edge Analytics Runtime.

The application provides services that can be extended to perform the following tasks:

• Ingest a pre-configured analytic template to the catalog for the supported runtime environments.• Create the analytic template and upload the developed analytics for the supported runtime

environments into the analytic catalog.• Search, sort, and filter analytic templates and deployments.• Configure analytic templates to add supporting documents, inputs, outputs, constants, and output

events.• Create and configure asset filters for defining asset applicability.• Add deployment configuration using an existing analytic template.• Orchestrate analytic data flows using existing analytic templates.• Apply the deployment configuration to applicable assets. Then map inputs and outputs to the tags in

the selected assets.• Schedule deployment jobs to automatically run in the configured runtime environment.

Architecture and ComponentsIn this section, you will learn about the functional architecture and various components within theApplication Analytics.

The different components of Application Analytics are shown in the following functional architecturediagram:

2 © 2020 General Electric Company

Page 7: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Application Analytics consists of the following primary components:

UI Widgets and AppsThe suite of visualization apps that enable you to add analytic templates to the framework's analyticcatalog, configure the templates, reuse templates in orchestrations and deployments, configure andschedule deployment jobs to run in the selected analytic runtime.

ServicesA suite of services that provides a standardized approach to upload analytic to deploy to varioussupported analytic runtimes. These services enable to get asset data for mapping your analyticinputs/outputs to applicable asset tags. They also provide the connectivity to the data providers suchas Predix and GE Digital APM times-series, and GE Digital APM asset data for running analytics onapplicable asset data and write back the generated outputs back into time-series and asset whenneeded.

RuntimeThe environment for running the deployed analytic jobs. The runtimes currently supported are Predix,Apache Spark, and SmartSignal.

© 2020 General Electric Company 3

Page 8: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Prerequisites for Getting Started with Application AnalyticsThe getting started instructions for Application Analytics assumes that the following preconditions aremet.

Tenancy Prerequisites

Before users can access the application analytics and services the system administrator must set up thetenant structure.

Your tenant must have access to the following services and features:

Predix Services

• predix-timeseries• predix-analytics-framework• predix-event-hub• predix-insights

Features

Note: Refer to the respective runtime pre-requisites sections for enabling features related to eachruntime.

• Alerts > Manage Alerts• Analysis > Analytics-Create and Analysis-View• Analytics

Following are the list of features for the supported runtime environments. You must select the onesthat apply to your tenant. Access to runtimes depends on the initial tenant setup:

◦ PowerDataFabric◦ Spark◦ SmartSignal◦ SOAR◦ AIR◦ Predix

• Asset > all• Audit Logs• Dashboard > Dashboard View, Update, Create• Templates > Alert Templates• Timeseries > all

Set Up the Analytics Functional UserBecause of security considerations, prior to accessing analytic modules and apps, the tenantadministrator must set up the analytics functional user. Without assigning the password for the functionaluser, the app modules and functionality will not be accessible.

Before You Begin

You have access to the newly setup tenant.

4 © 2020 General Electric Company

Page 9: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About This Task

You are required to do this the very first time you login. To set up the password, do the following:

Procedure

1. Login to your tenant with the tenant administrator credentials.2. Select Setup to access the Setup page.

Note: If you are setting up the password for the first time, you will be presented with a passwordsetup boxes in the Analytics section of the page.

3. Enter a new password, and then re-enter the password to confirm.4. Select Submit to save your changes.

Results

The analytics module and apps should be accessible within a few minutes after the functional user hasbeen set up.

Analytic Development PrerequisitesAs a data scientist or analytic developer, make sure you refer to the Developer section for analyticsamples and understanding the computation logic provided there.

You may also need the following:

• Access to Application Analytics samples in the GitHub repository https://github.build.ge.com/APM-Analytics/apm-analytics-samples. Contact support if you need access to thesample analytics on GitHub.

• Environment conducive to developing and testing analytics locally. The requirements may vary basedon the runtime and programming language used to code.

Table 1: Runtime Specific Reference Documentation

Runtime Developer Documentation Link

SmartSignal https://www.ge.com/digital/documentation/predix-apm/

IYzEzNTMxZTgtN2JjMC00NjZjLTg5OTUtYjJkOTRmNDk1YjY4.htm

l

Predix Analytics Runtime https://www.ge.com/digital/documentation/predix-services/

IYTYwZDgyZGUtOWI5Ny00OWY1LTllYWUtZWE0YTZlMjc3NjFl.h

tml#IYTYwZDgyZGUtOWI5Ny00OWY1LTllYWUtZWE0YTZlMjc3

NjFl

Spark Analytics https://www.ge.com/digital/documentation/opm/

Chunk238397185.html

© 2020 General Electric Company 5

Page 10: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

2WorkflowTopics:

• Analytics User TypicalWorkflow

6 © 2020 General Electric Company

Page 11: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Analytics User Typical WorkflowThe analytics user uses historical data to design and test analytics that identify important events in assetperformance data. Many queries are open-ended and require significant, manual data import. Workingclosely with analysts, they also collaborate with engineering and product design teams to developanalytics. They also develop analytics for use by plant operations in detecting anomalies.

The following diagram represents a typical workflow using Application Analytics. The table following thediagram lists the functional tasks.

Figure 1: Analytic User Typical Work flow

Table 2: Functional Tasks

Task Set Description

Manage Asset Filters (applicability) • Add, modify and delete asset filters for defining the assetapplicability for analytic.

Manage Analytic Templates • Add, modify, and delete analytic templates in the analyticcatalog.

• Configure analytic templates.

◦ Add supporting documentation.◦ Define the analytic inputs, constants and outputs.◦ Add an output event that matches an existing alert

template.

Manage Orchestrations • Add, modify, and delete analytic orchestrations and dataflows using existing analytic templates.

Manage Deployments • Add a deployment configuration.• Select assets for deployment.• Maps inputs and outputs definitions with tags and

attributes for deployment.• Modify constants values if needed.• Add iterations to analytic orchestration steps.

Schedule Deployments • Add data interval (sampling) and execution time range.• Schedule deployments to run once, on demand .• Schedule deployments to run automatically on a specified

date/time.• Review the deployment.

© 2020 General Electric Company 7

Page 12: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

3AlertsTopics:

• Generate Multiple Alerts

8 © 2020 General Electric Company

Page 13: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Generate Multiple Alerts

About This Task

To perform scalable alert ingestion without delay, you can generate multiple alerts at the same time foran analytic that is deployed to Predix working on an Apache Spark runtime environment. This topicdescribes how to enable this functionality for an analytic.

Before You Begin

If you have deployed analytics to generate alerts, redeploy the analytics.

Procedure

1. Access the configuration page of the analytic template.2. In the Output Events section, add the alert templates corresponding to the alerts that the analytic

can generate.3. Access the I/O Mapping page.4. In the Output Events section, select the alert templates corresponding to the alerts that you want to

generate.

Results

Multiple alerts are now generated for the selected alert templates.

© 2020 General Electric Company 9

Page 14: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

4Asset Filters and ConfigurationTopics:

• Analytics Configuration• Manage Asset Filters

10 © 2020 General Electric Company

Page 15: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Analytics Configuration

About Analytics ConfigurationConfiguration allows the user to define existing asset model attributes to use in asset filters.

The Analytics Configuration module allows you to view, add, modify, or delete attributes that are used inexisting asset filters. You need special access permission to add, modify, or delete existing asset modelattributes to the analytics configuration. You cannot modify or delete attributes configuration that havebeen used in an asset filter.

Important: You can only use system attributes and custom attributes that have been previously ingestedinto your tenant's asset model. Using non-existing model attributes in your configuration will not addthem to the asset model, nor will they show up for use in the asset filters as they do not exist.

What is Analytics Configuration?

During asset filter creation or modification, you can apply rules to define asset applicability based oncertain attributes. You should not define attributes that do not exist in your tenant's asset model. Forexample, you can define filter assets with all 7FA frame sizes except for assets with 7FA.05 as itsEquipment Code. In the context of asset filter, you can add rules to define the key value pairs for theseattributes. In this case, the rule with the IS operator can be applied to the attributes Frame Size andEquipment Code that have values 7FA and 7FA.05 respectively.

The Analytics Configuration module maintains the attributes configuration used in the asset filter rules.The rules are then used for retrieving results from the GE Digital APM Asset Model. Example 1 below,shows a rule that can be defined in the Asset section of the asset filter. This filter criteria Asset Type isGE90, where Asset Type is an Attribute and GE90 is the Value. Asset Type is an Attribute that must bedefined within the Analytics Configuration in order to be used within an Asset Filter.

Example 1:

Return assets where Asset Type is GE90

Note: Attributes defined within the Analytics Configuration must be identical to the attributes definedwithin the GE Digital APM Asset Model.

There are currently five columns present within the Configuration:

• Asset• Segment• Site• Enterprise• Custom Attributes

Important: Using custom attributes in asset filters is only available for tenants that have the featureenabled in the analytic_tenant_config table during tenant provisioning. Contact your tenant or systemadministrator if you want to configure asset filters with custom attributes.

Asset, Segment, Site, and Enterprise align with the asset business hierarchy in the GE Digital APM AssetModel. The Custom Attributes column maintains attributes that are not linked to the asset businesshierarchy. However, these attributes must be defined as part of your ingested asset model before usingthem in the Analytics Configuration.

© 2020 General Electric Company 11

Page 16: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Add Asset Filter Attributes

You can add asset filter attributes that are used to apply operational rules during analytic deployment foreach analytic level.

About This Task

The analytic levels are consistent with the asset model and include:

• Asset• Segment• Site• Enterprise

You can create an asset filter that contains specified attributes and their values. You can use asset filtersduring asset selection in deployment to determine the assets needed to deploy the analytic.

Procedure

1. In the module navigation menu, go to Analytics > Configuration.2. Select + Filter Attribute button in the desired level to add the attribute.3. Enter the attribute name.4. Select the blue check mark when finished to save your changes.

The new attribute is saved and becomes available within the Configuration.

Modify Asset Filter Attributes

You can modify asset filter attributes that are used to apply operational rules during analytic deploymentfor each analytic level.

About This Task

The analytic levels are consistent with the asset model and include:

• Asset• Segment• Site• Enterprise

You can modify the asset filter attributes and their values. You can use the configuration to filter assetsduring asset selection in deployment to determine the assets and attributes needed to deploy theanalytic.

Procedure

1. In the module navigation menu, go to Analytics > Configuration.2. Select the attribute to modify.3. Make necessary modifications to the attribute4. Select the blue check mark when finished to save your changes.

The modifications are saved in the Configuration.

12 © 2020 General Electric Company

Page 17: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Asset Filters

About Asset FiltersAsset Filters determine asset applicability when deploying an analytic.

During the configuration of an analytic deployment, users can define the asset(s) that the analytic willdeploy against. Applicability determines what assets within the asset model the analytic is intended toexecute on. This is performed through the use of Asset Filters, which captures an applicable group ofassets and/or types of assets. Asset Filters can be used by any number of analytic deployments that sharethe same asset applicability. For example, an analytic that detects high vibration anomalies on gasturbines would use an Asset Filter for all gas turbine type assets.

What is Applicability?

Asset Filters are saved GE Digital APM asset queries. These filters can be reused to define assetapplicability during analytic and orchestration deployments. The retrieval of assets based on Asset Filtershappens during deployment. Input and output mappings defined using the Map for all assets option willbe applied to all assets resolved from the selected Asset Filter.

For Orchestration deployment, the Asset Filter is resolved once and used for each step of theorchestration. Asset Filters are created using one or more rules within each asset deployment hierarchylevel, that is, asset, segment, site, and enterprise.

Asset Filters allows you to add rules at any level in the asset business hierarchy when creating a query.Asset Filters allow adding multiple values for a single rule. It also allows the Is not operation.

Example1:

Return assets where Asset Type Is GE90 OR CF34 AND position Is not 1

Example2:

Return Assets where Site Name Is Doha AND Train Is 5 AND Asset Type Is Wind Turbine OR Gas Turbine OR Steam Turbine

Example3:

Return Sites where Asset Type Is Wind Turbine AND Country Is USA AND Site Type Is Wind farm

Asset Filters

You can use filter sets to create Asset Filters that can be used for asset applicability during thedeployment of an analytic.

You can access, but cannot modify the Asset Filter if it has been used in a deployment. You need deletepermissions to delete an existing Asset Filter. Make sure you are aware of any applicability usage prior todeleting an Asset Filter. You cannot delete an Asset Filter that has been used in a deployment.

© 2020 General Electric Company 13

Page 18: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

When the Asset Filter is used in a deployment, the analytic is executed against all assets that meet thefilter criteria. In addition, during I/O mapping, you can only access and map tags that are associated withassets that meet the filter criteria.

The following operators can be used within each rule:

• Is• Is not

Query Logic

The following query logic is used to retrieve query results from the asset model:

• When multiple values are entered within a rule, the OR condition is applied. For example, for a singlerule where Asset Type is GE90, CF34 is defined, the query results in all asset types matchingGE90 or CF34.

• When multiple rules are entered, the AND condition is applied. For example, if you define a rule -Asset Type Is Wind Turbine and another rule - Country Is USA, the query should resultin asset type that are Wind Turbines located in USA.

Note: You cannot apply rules at a lower hierarchy level than the asset deployment hierarchy level of theAsset Filter.

GE Digital APM provides API endpoints to add additional query fields. For more information on using theAPIs, contact GE Global Support.

Create Asset Filters

You can create asset filters that are used to determine asset applicability while deploying an analytic.

Before You BeginYou must have the required access permissions to perform this task.

About This Task

You can create an asset filter that contains specified attributes and their values. You can use asset filtersduring asset selection in deployment to determine the assets needed to deploy the analytic.

Important: Using custom attributes in asset filters is only available for tenants that have the featureenabled in the analytic_tenant_config table during tenant provisioning. Contact your tenant or systemadministrator if you want to configure asset filters with custom attributes.

Procedure

1. In the module navigation menu, go to Analytics > Asset Filters.

2. Select on the page heading.3. In the workspace of the Asset Filters page, enter a unique name for the asset filter in the NAME box.4. Enter a description in the DESCRIPTION box.5. Select one of the options in the ASSET HIERARCHY DEPLOYMENT LEVEL drop-down list box. The

default selection is Asset.6. In the respective section, select + Add Rule to add a new attribute filter.

Note: A disabled + Add Rule box indicates that all available attributes have already been used in theasset filter.

a) In the ATTRIBUTE box, select an attribute value and then select a value in the CONDITION box.b) In the VALUE box, enter one or more values as the search keywords for the filter. Use commas to

separate multiple values.

14 © 2020 General Electric Company

Page 19: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Each attribute appears in the filter set. To remove an attribute, select at the end of the row forthe attribute.

7. Add additional filters as needed. For each filter combination, select or enter the parameters and valuesneeded to complete the string.For example, to include all asset names that start with GE-9 in the model, but exclude GE-901, createthis combination:

a. In the Assets section, select + Add Rule.b. In the ATTRIBUTE box, select Asset Name, and then select Is in the CONDITION box.c. Enter GE-9* to find all asset names that start with GE-9.d. In the Assets section, select + Add Rule to add another rule.e. In the ATTRIBUTE box, select Asset Name, and then select Is not in the CONDITION box.f. Enter GE-901 to exclude all asset names that start match GE-901.

Note: You must enter a value for each attribute field. You can use wildcards in search.

8. Select Create.

Attention: If you reset prior to saving, all changes and additions for the asset filter areremoved.

The asset filter is saved and becomes available in the pane for selection during deployment.

Next Steps

Before you use the newly added asset filter in a deployment, preview the filter results and make anyrequired modifications.

Access and Modify Asset Filters

You can preview the filter results and make any required modifications to an existing asset filter.

Before You Begin

You must have the required access permissions to perform this task.

About This Task

Make sure that you are aware of any applicability usage before modifying an asset filter. Consider creatinga new asset filter instead.

Note: You cannot modify an asset filter if the feature analytic_tenant_config table is enabled duringprovisioning.

Note: You cannot modify an asset filter if it has been used in a deployment.

Note: You cannot apply rules at a lower hierarchy level than the asset deployment hierarchy level of theasset filter.

Procedure

1. In the module navigation menu, go to Analytics > Asset Filters.

2. In the Search pane, select , and then enter characters in the box to find an asset filter.Suggestions narrow as you enter characters.

3. Select an asset filter to open it in the workspace.

4. Select Preview to access the Asset Filter Sample Result window.

© 2020 General Electric Company 15

Page 20: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

5. Make any necessary modifications to name, description, and rules. You cannot modify the values in theASSET HIERARCHY DEPLOYMENT LEVEL box.

6. Repeat the preview and modification steps till you get the expected results from the asset filter.7. Select Save Changes to save the asset filter or Cancel to cancel the modifications.

Delete Asset Filters

You can delete an asset filter that is not being used by a deployment.

Before You Begin

You must have the required access permissions to perform this task. Deletion is currently disabled intenants that have

About This Task

You can delete an asset filter that contains specified attributes and their values. Deleting an asset filtercannot be undone.

Note: You cannot delete an asset filter if the feature analytic_tenant_config table is enabled duringprovisioning.

Note: You cannot delete an asset filter if it has been used in a deployment.

Procedure

1. In the module navigation menu, go to Analytics > Asset Filters.

2. In the Search pane, select , and then enter characters in the box to find an asset filter.Suggestions narrow as you enter characters.

3. Select an asset filter to open it in the workspace.4. Select Delete.

Results

The asset filter is removed from the pane and is no longer available for use during deployment.

16 © 2020 General Electric Company

Page 21: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

5Analytic TemplatesTopics:

• Manage Analytic Templates• Configuration

© 2020 General Electric Company 17

Page 22: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Analytic Templates

About Analytics TemplatesIn this section, you will learn about analytic templates and the analytic catalog. Analytics generate data tohelp you manage your assets. Application Analytics enables you to build analytic templates and uploadanalytics to the analytic catalog.

Analytics implement mathematical models to provide intelligence about your assets, helping you managethem more efficiently. For example, you can use analytics to trigger alerts, generate maintenanceadvisories, and produce performance statistics.

Application Analytics enables you to create an analytic template that you can use in multipledeployments for the specified runtime. By employing a template, you can reuse the same analytic todeploy to different assets using asset applicability, and create unique input-output (I/O) tag mapping perdeployment. You can also vary the schedule per deployment. The framework expects unique name perdeployment.

The analytic catalog groups the tasks necessary to successfully configure and deploy an analytic.

Add an Analytic Template

Add the analytic packaged implementation to the Application analytic catalog to be deployed to a specificanalytic runtime for running the analytic.

Before You Begin

Be sure that the packaged implementation containing the code for your analytic and any supporting files,such as instructions for deployment, are available to upload.

Note: If you are using the Import from file option to ingest analytics, and if your analyticsimplementation contains more than 200 entries for inputs definitions, constants, and output definitions,make sure that you set the isReadyOnly key to true. This is necessary to avoid latency issues whileloading multiple definitions on the UI.

About This Task

The system validates the Name and Analytic File fields before you can proceed. You will need to enter orselect values based on the selected runtime type.

Depending on the type of analytic implementation the configuration options vary for adding the templateto the catalog. The analytic runtime environments supported for your tenant vary based on the tenantconfiguration.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.

2. Select in the page heading.3. Add the analytic template using one of the following methods:

• Select the Import from file box. To ingest the complete analytic template, you must have the validanalytics ingestion .zip file ready. Refer to the topic Understand Analytic Ingestion in thereference section.

a) Select the Import from file box.

18 © 2020 General Electric Company

Page 23: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

b) Browse to the location of the analytic ingestion .zip file and select the file.c) Select Open.

• Use the fields in the New Analytic page to configure the template.

a) In the New Analytic page, select a value in the Runtime Type box.b) Select a value in the Data Frame Type box.

The values in the box may vary, depending on the value that you selected in the Runtime Type box.This value determines the way the Application Analytics expects the input.json file format forthe analytic algorithm. It also determines the format in which Application Analytics creates thecorresponding output.json and PortToField.json files. For example, if you select theShared Time Stamp option, then the input.json file must give the values for all sharedtimestamps within the timeseries data. Otherwise, if you select the Per Port Time Stamp option,then the input.json must provide per-port values for inputs and output.json will have per-port values created for outputs.

c) Enter or select values in the Template Details section of the page.4. Select Submit.

Results

The template is saved to the catalog. After the analytic logic is successfully uploaded, the status changesto Completed.

Next Steps

Configure and build the analytic template after adding it to the catalog.

Related Concepts

About Analytic Ingestion Services on page 105

Application Analytics provides specific REST API services to upload or ingest your analytic logic,template definitions, and deployments simultaneously.

Related Reference

Analytic Template Fields based on Runtime on page 19

List of analytic template configuration requirements for different runtimes.

Analytic Template Fields based on Runtime

List of analytic template configuration requirements for different runtimes.

Field Requirements Based On Runtime

Refer to the table below to see the required and optional data entry fields based on the selected analyticruntime for the analytic template. Availability of a specific runtime is dependent on the tenantconfiguration.

Field Name Description

Predix Spark PredixEdge

SmartSignal

AIR SOAR PowerFabric

Data Frame

Types

This value

determines

the way the

Application

• Shared

Time

Stamp

N/A N/A N/A N/A N/A N/A

© 2020 General Electric Company 19

Page 24: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Field Name Description

Predix Spark PredixEdge

SmartSignal

AIR SOAR PowerFabric

Analytics

expects the

input.json file

format for

the analytic

algorithm. It

also

determines

the format

in which

Application

Analytics

creates the

correspondi

ng

output.json and

PortToField.json files.

• Per Port

Time

Stamp

Name A unique

name for

the analytic

template.

Required Required Required Required Required Required Required

Owner The name of

the analytic

owner,

typically the

developer of

the analytic

logic.

Required Required Required Required Required Required Required

Analytic

Type

The

language

used to

develop the

analytic.

• Java

• Python

• Matlab

• Java

• Python

Edge

Analytic

Editor

Smartsignal • Java

• Python

• Java

• Python

• Edge

Analytic

Editor

CCAP

(Version:

3.1)

Type

Version

The version

for the

selected

language.

The type

version is

updated

based on

the analytic

Required Required Required Required As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

20 © 2020 General Electric Company

Page 25: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Field Name Description

Predix Spark PredixEdge

SmartSignal

AIR SOAR PowerFabric

type

selected.

Analytic FileSelect

to upload

the

packaged

analytic

code.

Note: The

uploaded

file must be

in .zip or .jar

format.

Required Required Required Required Required Required Required

Analytic

Version

Version of

the

uploaded

analytic

implementa

tion.

Required Required Required Required Required Required Required

Analytic

Version

Update

The

information

specific to

what is new

or changed

in this

version of

the

uploaded

analytic

package.

Optional Optional Optional Optional Optional Optional Optional

Primary

Category

The main

category tag

used for the

analytic.

This tag

allows the

analytic to

be classified

under the

designated

category for

further

discovery in

the catalog.

Enter a

value that

best

Required Required;

Must be

Forecasting for

forecasting

and What-If

analytics.

Required Required As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

© 2020 General Electric Company 21

Page 26: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Field Name Description

Predix Spark PredixEdge

SmartSignal

AIR SOAR PowerFabric

matches the

type of

analytic you

are creating.

For

example, if

monitoring

a turbine,

you can

select

Monitoring.

Secondary

Category

The

secondary

category tag

used for the

analytic.

This tag

allows the

analytic to

be classified

under the

designated

category,

allowing you

to further

refine where

the analytic

is applied.

Optional Optional Optional Optional As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

As defined

in the

runtime

configuratio

n

Note: You can also define custom runtime, or Bring Your Own Runtime (BYOR), which is a configurationthat you define. You must work with the Application Analytics team to define the contract for the runtimeand enable it on your tenant. Customization must be done during tenancy setup. This is similar toenabling other default runtime configurations.

22 © 2020 General Electric Company

Page 27: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 3: Analytic Version Support Based On Runtime

Runtime Multiple Version Support? Instructions To Upload And Save ANew Version

Predix YES To save a new version of the analytic (.zip

or .jar) file, upload a new file, enter the

new Analytic Version number, and then

enter the Analytic Version Update

description. The latest version of the

analytic artifact will be used to run all

future deployments. Existing

deployments continue to use the

currently assigned version unless you

explicitly redeploy them to run with the

new version.

Spark NO Uploading multiple versions of the

analytic to an existing template entry in

the catalog is not currently supported.

SmartSignal YES To save a new version of the analytic (.zip

or .jar) file, upload a new file, enter the

new Analytic Version number, and then

enter the Analytic Version Update

description. The latest version of the

analytic artifact will be used to run all

future deployments. Existing

deployments continue to use the

currently assigned version unless you

explicitly redeploy them to run with the

new version.

© 2020 General Electric Company 23

Page 28: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Runtime Multiple Version Support? Instructions To Upload And Save ANew Version

Predix Edge NO Uploading multiple versions of the

analytic to an existing template entry in

the catalog is not currently supported.

To save a new version of the analytic file,

do one of the following:

• From the Edge Analytics Editor,

publish a new version of your analytic

to the Analytics Catalog.

• Manually upload a new .eae file into

the Analytics Catalog.

Enter the new Analytic Version number,

and then enter the Analytic Version

Update description. The latest version of

the analytic artifact will be used to run all

future deployments. Existing

deployments continue to use the

currently assigned version unless you

explicitly redeploy them to run with the

new version.

AIR, SOAR, and Power Datafabric Varies Depends on the runtime definition and

configuration.

Supported Data Sources

Data Source Predix Spark Edge SmartSignal AIR SOAR Power Fabric

GE Digital

APM Time

Series

Yes Yes Yes Yes Yes No Yes

GE Digital

APM Asset

Attributes

Yes Yes Yes No No No Yes

Constant Yes Yes Yes Yes Yes No Yes

Blobstore No Yes No No No No No

Any No No No No Yes Yes No

Postgres No Yes No No No No No

Predix

Timeseries

No No No No No No Yes

24 © 2020 General Electric Company

Page 29: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Modify an Analytic Template

Use the modify functionality to modify existing Analytic Catalog entries.

About This Task

You cannot modify the runtime type for an existing analytic template in the catalog. You may be able tomodify only some fields for the selected analytic depending on its runtime type. You can see values indisabled fields but cannot modify them.

Make sure you check all existing deployments that use the analytics template before you modify it. If yourmodified artifact has analytic implementation that is significantly different from the previously uploadedartifact in the catalog, consider creating a new analytic to add to the catalog instead of updating theexisting one.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.2. In the Search pane, locate and select an analytics to access it.

3. In the workspace, select to access the analytic template for modification.4. In the Analytics Template page, you can modify values in some of the existing boxes.5. When you have completed your modifications, select Submit.

Results

After selecting Submit, the page reflects your changes to the analytic.

Delete an Analytic Template

You can delete an analytic template in the analytic catalog.

About This Task

Before you delete an analytic template in the analytic catalog, make sure that it is not being used inorchestrations or deployments. Deleting an analytic deletes its deployments. This cannot be undone.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.2. In the Search pane, locate and select an analytics to access it.3. In the workspace of the Analytics Template page, select Delete.4. In the Delete the Analytic window, select Delete.

Note: This cannot be undone.

© 2020 General Electric Company 25

Page 30: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Configuration

Configure the Analytics TemplateThis task enables you to record inputs, constants, outputs, output events, and supporting documentationin an analytic template.

Before You Begin

• Make sure you have uploaded the analytics to the catalog.• If your analytics are expected to create event based alerts, you must have the alert template defined

prior to referencing it in the output events of the analytic template.

About This Task

Use the analytic template to add information required to deploy the analytics. For example, you can add adocument explaining how the inputs and outputs for the analytics were determined.

After the analytics have been created and added to the catalog, you can build the analytics template. Thetemplate definitions must directly correlate to the developed analytics. For example, if your analyticsexpects certain inputs, constants, or outputs, make sure you define them in your template.

The Save and Cancel buttons appear after you have updated any information. You can select thesebuttons at any time. Selecting the button saves or cancels changes from the point of last your save orcancel operation.

Searching for an alert template in the Output Events section provides suggestions as you type.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.

2. In the Search pane, locate and select an analytics to access it.

3. Optional: In the Analytic Template page, select the version link below the analytics name, thenselect the Make Current button next to the version in the list that appears to change the analyticsversion.

The link is available only if you have added more than one version for the same analytics to thecatalog. The defined runtime will use the analytics current version for running all future deployments.For example, if you have two uploaded versions, 1.0.0 and 2.0.0, to DowntimeForTurbinesanalytics in the catalog, you can select either 1.0.0 or 2.0.0 to run all future deployments ofDowntimeForTurbines.

4. Select the Analytic Template tab.

5. Optional: Select in the AGGREGATE MAPPING subsection to allow multiple tags to bemapped to a single port.

The button changes to selected ( ).

6. Hover over Select to add a description in the Description section, then select to enter adescription for the analytics.

The analytics description can state its purpose, and can include types of trends or patterns. Forexample, you can add Use the Turbine analytics to monitor unplanned turbinedowntime.

26 © 2020 General Electric Company

Page 31: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

7. In the Supporting Files section, select to add supporting files to the analytics template.

8. As needed, enter values in the Input Definition, Constants, and Output Definition sections.

9. In the Output Events section, select the Name box and begin entering the Alert Template name. GEDigital APM provides suggestions as you enter characters.

The Alert Template name is validated and an error message appears if it does not exist. You can mapmultiple output events for a given analytic template.

10. Verify your entries. Select Save to save changes to the template. Select Cancel to discard all changes.You must confirm or cancel your choice in the window that appears.

Results

After a successful save operation, all changes made to the analytics template are saved in the analyticcatalog.

Next Steps

Add deployments for the analytics.

Upload Supporting Files for the Analytic Template

You can upload supporting documentation for an analytic in the analytic template.

Before You Begin

Make sure you have created a catalog entry for the analytic.

About This Task

Supporting documents provide more information on the analytic, such as process-information diagramsor plant layouts. You can upload the following document types:

• image files (.jpg, .jpeg, .png, .bmp, .tif)• Text type files (.doc, .docx, .xls, .xslx, .csv, .ppt, .pptx, .txt)• .pdf files

You cannot add folders or .exe, .bat, .cmd, .msi, .bin, .zip, or .rar files.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.

2. In the Search pane, locate and select an analytic in the list to access it.

3. Select the Analytic Template tab.

4. Select in the Supporting Files section.

5. In the Upload New File window, select the File box to locate the file, enter a description in itsDescription box, and then select Submit.The uploaded file is listed in the SUPPORTING FILES subsection. Repeat the procedure to uploadmultiple files, if needed.

Modify a Supporting File Description

You can modify the description for a supporting file that has been uploaded.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.

© 2020 General Electric Company 27

Page 32: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

2. In the Search pane, locate and select an analytic in the list to access it.3. Select the Analytics Template tab.4. Hover over the Description column of the supporting file listed in the SUPPORTING FILES

subsection, and then select the description or select .5. Modify the description in the box, and then select outside the box.6. Select Save to save changes to the analytic template. Select Cancel to discard all changes. You must

confirm or cancel your choice in the window that appears.

Download Template Logs from the Analytic Template

You can download a .zip file of logs containing all analytic deployments for the current version of theanalytic template.

Before You Begin

You have access to analytic templates and deployments. You have already deployed the analytic and ran it

at least once in the designated runtime. The button will not be visible if there are nocurrent deployments for the selected analytic template.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.2. In the Search pane, locate and select an analytic in the list to access it.3. Select the Analytic Template tab.

4. In the page heading, select to download the logs as a .zip file to your device.

Results

The .zip file contains the deployment log file.

Tip: Search for the following keywords in the log file. These keywords help you determine if the analyticdeployment ran successfully.

AnalyticInputJson shows the analytic inputs. Data that the analytic received from mapped input tags andattributes.

AnalyticInputJson:{ "data":{ "time_series":{ "time_stamp":[ 1512184000000 ], "temp_today":[ 90.0 ] }, "constants":{ "threshold":1.0 } }}

28 © 2020 General Electric Company

Page 33: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

AnalyticResponseJson shows the response or the analytic output . The response with outputs indicatesthat the analytic deployment ran successfully and data was written to the mapped outputs.

AnalyticResponseJson:{ "alerts":{ "commonKeyValueStore":{

}, "commonSensors":[

], "commonScanGroupData":{ "threshold":1.0 }, "events":[ { "date":1512184000000, "keyValueStore":{

}, "sensors":[ "temp_today" ], "scanGroupData":{ "scores":90.0 } } ] }, "time_series":{ "time_stamp":[ 1512184000000 ], "mean":[ 90.0 ], "deviation":[ 1.0 ] }}

Example

The following is a sample of the log file:

Created app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98fUpdated app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f ({"route"=>"379cc9ac-c29d-4957-9462-c29c12f17534", :verb=>"add", :relation=>"routes", :related_guid=>"379cc9ac-c29d-4957-9462-c29c12f17534"})Updated app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f ({"state"=>"STOPPED"})Updated app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f ({"environment_json"=>"PRIVATE DATA HIDDEN"})Updated app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f ({"state"=>"STARTED"})Downloading java_buildpack...

© 2020 General Electric Company 29

Page 34: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Downloaded java_buildpackCreating containerSuccessfully created containerDownloading app package...Downloaded app package (24.1M)Staging...-----> Java Buildpack Version: v3.9 | https://github.com/cloudfoundry/java-buildpack.git#b050954-----> Downloading Open Jdk JRE 1.8.0_101 from https://java-buildpack.cloudfoundry.org/openjdk/trusty/x86_64/openjdk-1.8.0_101.tar.gz (2.0s) Expanding Open Jdk JRE to .java-buildpack/open_jdk_jre (1.1s)-----> Downloading Open JDK Like Memory Calculator 2.0.2_RELEASE from https://java-buildpack.cloudfoundry.org/memory-calculator/trusty/x86_64/memory-calculator-2.0.2_RELEASE.tar.gz (0.0s) Memory Settings: -Xms681574K -XX:MetaspaceSize=104857K -Xss349K -Xmx681574K -XX:MaxMetaspaceSize=104857K-----> Downloading Spring Auto Reconfiguration 1.12.0_RELEASE from https://java-buildpack.cloudfoundry.org/auto-reconfiguration/auto-reconfiguration-1.12.0_RELEASE.jar (0.1s)Exit status 0Staging completeUploading droplet, build artifacts cache...Uploading build artifacts cache...Uploading droplet...Uploaded build artifacts cache (44.9M)Creating droplet for app with guid 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98fUploaded droplet (66.6M)Uploading completeStopping instance 002861de-2877-41e5-912e-eccee53dfe9dDestroying containerCreating containerSuccessfully destroyed containerSuccessfully created containerStarting health monitoring of container04:34:44,775 |-WARN in ch.qos.logback.core.rolling.RollingFileAppender[null] - For more information, please visit http://logback.qos.ch/codes.html#rfa_no_tp04:34:44,775 |-INFO in ch.qos.logback.classic.joran.action.RootLoggerAction - Setting level of ROOT logger to INFO04:34:44,776 |-INFO in ch.qos.logback.core.joran.action.AppenderRefAction - Attaching appender named [consoleAppender] to Logger[ROOT]04:34:44,776 |-INFO in ch.qos.logback.core.joran.action.AppenderRefAction - Attaching appender named [FILE] to Logger[ROOT]04:34:44,777 |-ERROR in ch.qos.logback.core.joran.action.AppenderRefAction - Could not find an appender named [LOGSTASHLOG]. Did you define it below instead of above in the configuration file?04:34:44,777 |-ERROR in ch.qos.logback.core.joran.action.AppenderRefAction - See

30 © 2020 General Electric Company

Page 35: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

http://logback.qos.ch/codes.html#appender_order for more details.04:34:44,777 |-INFO in ch.qos.logback.classic.joran.action.ConfigurationAction - End of configuration.04:34:44,777 |-INFO in ch.qos.logback.classic.joran.JoranConfigurator@542055f3 - Registering current configuration as safe fallback point2019-03-07 04:34:47,095 3555 px_correlation_id= px_zone_id= px_service= px_user_name= [main] INFO o.c.r.s.CloudPropertySourceApplicationContextInitializer - Adding 'cloud' PropertySource to ApplicationContext2019-03-07 04:34:50,859 7319 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO c.g.p.a.j.m.AnalyticEnvironmentContext - Application Name is : 76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c22019-03-07 04:34:50,861 7321 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO c.g.p.a.j.m.AnalyticEnvironmentContext - Application Name is : 76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c22019-03-07 04:34:50,868 7328 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO c.g.p.a.j.m.AnalyticEnvironmentContext - Application Name is : 76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c22019-03-07 04:34:51,341 7801 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO org.apache.cxf.endpoint.ServerImpl - Setting the server's publish address to be /2019-03-07 04:34:51,632 8092 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.m.m.a.RequestMappingHandlerAdapter - Looking for @ControllerAdvice: org.springframework.boot.context.embedded.AnnotationConfigEmbeddedWebApplicationContext@718a62d3: startup date [Thu Mar 07 04:34:47 UTC 2019]; root of context hierarchy2019-03-07 04:34:51,794 8254 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.m.m.a.RequestMappingHandlerMapping - Mapped "{[/error],methods=[],params=[],headers=[],consumes=[],produces=[text/html],custom=[]}" onto public org.springframework.web.servlet.ModelAndView org.springframework.boot.autoconfigure.web.BasicErrorController.errorHtml(javax.servlet.http.HttpServletRequest)2019-03-07 04:34:51,794 8254 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.m.m.a.RequestMappingHandlerMapping - Mapped "{[/error],methods=[],params=[],headers=[],consumes=[],produces=[],custom=[]}" onto public org.springframework.http.ResponseEntity<java.util.Map<java.lang.String, java.lang.Object>> org.springframework.boot.autoconfigure.web.BasicErrorController.error(javax.servlet.http.HttpServletRequest)2019-03-07 04:34:51,822 8282 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.h.SimpleUrlHandlerMapping - Mapped URL

© 2020 General Electric Company 31

Page 36: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

path [/webjars/**] onto handler of type [class org.springframework.web.servlet.resource.ResourceHttpRequestHandler]2019-03-07 04:34:51,822 8282 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.h.SimpleUrlHandlerMapping - Mapped URL path [/**] onto handler of type [class org.springframework.web.servlet.resource.ResourceHttpRequestHandler]2019-03-07 04:34:51,865 8325 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.w.s.h.SimpleUrlHandlerMapping - Mapped URL path [/**/favicon.ico] onto handler of type [class org.springframework.web.servlet.resource.ResourceHttpRequestHandler]2019-03-07 04:34:52,030 8490 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.j.e.a.AnnotationMBeanExporter - Registering beans for JMX exposure on startup2019-03-07 04:34:52,035 8495 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.c.s.DefaultLifecycleProcessor - Starting beans in phase 21474836472019-03-07 04:34:52,293 8753 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.a.coyote.http11.Http11NioProtocol - Initializing ProtocolHandler ["http-nio-8080"]2019-03-07 04:34:52,304 8764 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.a.coyote.http11.Http11NioProtocol - Starting ProtocolHandler ["http-nio-8080"]2019-03-07 04:34:52,310 8770 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.a.tomcat.util.net.NioSelectorPool - Using a shared selector for servlet write/read2019-03-07 04:34:52,339 8799 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO o.s.b.c.e.t.TomcatEmbeddedServletContainer - Tomcat started on port(s): 8080 (http)2019-03-07 04:34:52,341 8801 px_correlation_id=na px_zone_id=na px_service=analytic px_user_name= [main] INFO c.g.p.a.j.JavaBasedAnalyticApplication - Started JavaBasedAnalyticApplication in 7.096 seconds (JVM running for 9.999)Container became healthy2019-03-07 04:34:57,377 13837 px_correlation_id=6fed79c1-53b8-4456-bd8f-20d1459f2a18 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [http-nio-8080-exec-2] INFO c.g.p.a.j.r.c.AnalyticsEngineServiceClientImpl - Started async execution client2019-03-07 04:34:57,428 13888 px_correlation_id=6fed79c1-53b8-4456-bd8f-20d1459f2a18 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [http-nio-8080-exec-2] INFO c.g.p.a.j.r.c.AnalyticsEngineServiceClientImpl - AsyncExecutionRequestUri=https://predix-analytics-engine-release.run.aws-usw02-pr.ice.predix.io/api/v1/analytic/execution/async payload={}429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:34:58.413+0000]

32 © 2020 General Electric Company

Page 37: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"GET /api/v1/analytic/healthcheck HTTP/1.1" 200 0 0 "-" "Java/1.8.0_101" "10.72.11.19:60722" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.11.19" x_forwarded_proto:"https" vcap_request_id:"f8c2536c-f42c-4d96-7b70-481714480a85" response_time:0.007099449 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"b964cf2b87970abd" x_b3_spanid:"b964cf2b87970abd" x_b3_parentspanid:"-"

2019-03-07 04:34:58,936 15396 px_correlation_id=6fed79c1-53b8-4456-bd8f-20d1459f2a18 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [http-nio-8080-exec-6] INFO c.g.p.a.j.m.AnalyticEnvironmentContext - Application Name is : 76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c22019-03-07 04:34:58,936 15396 px_correlation_id=6fed79c1-53b8-4456-bd8f-20d1459f2a18 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [http-nio-8080-exec-6] 2019-03-07 04:40:05,724 322184 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [http-nio-8080-exec-2] INFO c.g.p.a.j.m.AnalyticEnvironmentContext - Application Id is : 1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f2019-03-07 04:40:05,729 322189 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO c.g.p.a.j.m.AnalyticAsyncExecutionReceiver - Received analytic execution message: com.ge.predix.analytics.common.dto.AnalyticEngineRequest@1e118338[tenantId=76ebcc0c-0cdd-4b7c-b53e-132383197760, analyticId=429fc52c-f855-4b81-8d7c-ce50e68505c2, analyticExecutionRequestId=04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0, readUrl=https://predix-analytics-engine-release.run.aws-usw02-pr.ice.predix.io/api/v1/analytic/execution/read/04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0, writeUrl=https://predix-analytics-engine-release.run.aws-usw02-pr.ice.predix.io/api/v1/analytic/execution/write/04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0, notificationUrl=https://predix-analytics-engine-release.run.aws-usw02-pr.ice.predix.io/api/v1/analytic/execution/notification/04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0, oAuthToken=eyJhbGciOiJSUzI1NiIsImtpZCI6ImxlZ2FjeS10b2tlbi1rZXkiLCJ0eXAiOiJKV1QifQ.eyJqdGkiOiI1MWExZmFlOWYzY2E0OTkxOTY1ZjEwZmZkMGVjNjZjMyIsInN1YiI6IjFqMjVsdjFoNnNtbWo0aDB5NDNubG8yY2txeHNpbnIwNXhmbGlyIiwic2NvcGUiOlsiYW5hbHl0aWNzLnpvbmVzLjc2ZWJjYzBjLTBjZGQtNGI3Yy1iNTNlLTEzMjM4MzE5Nzc2MC51c2VyIiwidGltZXNlcmllcy56b25lcy4zYjQwMzIxMC03NThlLTRhMDEtYjk3Yi03ZGRkNjVjYjMyMjUuaW5nZXN0IiwidGltZXNlcmllcy56b25lcy4zYjQwMzIxMC03NThlLTRhMDEtYjk3Yi03ZGRkNjVjYjMyMjUucXVlcnkiLCJ0aW1lc2VyaWVzLnpvbmVzLjNiNDAzMjEwLTc1OGUtNGEwMS1iOTd

© 2020 General Electric Company 33

Page 38: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

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.CUVVsIYgPkknCaLhsabraWNXT6m1Yxb4QhgNcQwjMw5ihTYoEIBVVZ4hrB_h7AL5zE6U3F5QYg4efQNqTDyhvHQutL801SRrwZZnD7ljnr1imGFo9IFtwMgWWyFaBOIQFgxBKFq9uEKhP3QsXZKkpX6w4SRXDH4Sl2wslMFftELZvZ3ysxgt6jCmO7oesRU-Niirb_3dVViwKLXqacokH03I2uyAJWD5JnuYEovmdNnINmZHGxdImUDTVhOr50yr1e_glWX6k6VoAW-6gMzJXTXXBWUODD1_DWgjTR7Aq7iK85f7zELHoDFM7CWxFFGTnOpmdOGCotxW5sbY6WDR3g, clientType=null, correlationId=ef4590bb-0e70-47e3-9e32-25f8f299b9e8, highCapacityTenant=<null>, data=<null>]2019-03-07 04:40:05,729 322189 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO c.g.p.a.j.m.AnalyticAsyncExecutionReceiver - Received analytic execution message - analytic id 429fc52c-f855-4b81-8d7c-ce50e68505c2, analytic request id 04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:40:05.724+0000] "POST /api/v1/analytic/execution/engine HTTP/1.1" 200 2558 0 "-" "Apache CXF 3.1.6" "10.72.1.204:17904" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.1.204" x_forwarded_proto:"https" vcap_request_id:"443afad1-e806-40d4-5f53-cfb919304dda" response_time:0.007693907 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"ca47f777dafdd0cc" x_b3_spanid:"ca47f777dafdd0cc" x_b3_parentspanid:"-"

2019-03-07 04:40:06,010 322470 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO com.ge.apm.analytics.ComputeStats - AnalyticInputJson: {"data":{"time_series":{"time_stamp":[1512184000000],"temp_today":[90.0]},"constants":{"threshold":1.0}}}2019-03-07 04:40:06,022 322482 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO com.ge.apm.analytics.ComputeStats - AnalyticResponseJson: {"alerts":{"commonKeyValueStore":{},"commonSensors":[],"commonScanGroupData":

34 © 2020 General Electric Company

Page 39: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

{"threshold":1.0},"events":[{"date":1512184000000,"keyValueStore":{},"sensors":["temp_today"],"scanGroupData":{"scores":90.0}}]},"time_series":{"time_stamp":[1512184000000],"mean":[90.0],"deviation":[1.0]}}2019-03-07 04:40:06,023 322483 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO c.g.p.a.j.execution.AESNotification - VcapApplication Configuration is {"application_id":"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f","application_name":"76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c2","application_uris":["429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io"],"application_version":"34a97419-9054-4005-8eae-bc5ca6ca22f4","cf_api":"https://api.system.aws-usw02-pr.ice.predix.io","host":"0.0.0.0","instance_id":"a50ba3f2-1ce8-49a2-459e-d0d4","instance_index":0,"limits":{"disk":2048,"fds":16384,"mem":1024},"name":"76ebcc0c-0cdd-4b7c-b53e-132383197760-429fc52c-f855-4b81-8d7c-ce50e68505c2","port":8080,"space_id":"9b30574b-23a5-4449-b0c7-26f16d1a0bfe","space_name":"76ebcc0c-0cdd-4b7c-b53e-132383197760","uris":["429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io"],"version":"34a97419-9054-4005-8eae-bc5ca6ca22f4"}2019-03-07 04:40:06,024 322484 px_correlation_id=ef4590bb-0e70-47e3-9e32-25f8f299b9e8 px_zone_id=76ebcc0c-0cdd-4b7c-b53e-132383197760 px_service=analytic px_user_name= [SimpleAsyncTaskExecutor-4] INFO c.g.p.a.j.execution.AESNotification - Configured memory is - 1024429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:40:06.388+0000] "GET /api/v1/analytic/healthcheck HTTP/1.1" 200 0 0 "-" "Java/1.8.0_101" "10.72.1.83:26154" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.1.83" x_forwarded_proto:"https" vcap_request_id:"22128387-24e0-4913-4538-d24208549e09" response_time:0.004521878 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"81b5d6273d102bb7" x_b3_spanid:"81b5d6273d102bb7" x_b3_parentspanid:"-"

429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:40:09.216+0000] "GET /api/v1/analytic/healthcheck HTTP/1.1" 200 0 0 "-" "Apache CXF 3.1.6" "10.72.1.204:27982" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.1.204" x_forwarded_proto:"https" vcap_request_id:"989e6a60-dd68-4c05-7a69-64ec14ce73be" response_time:0.002774438 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"08aa6454e6af589e" x_b3_spanid:"08aa6454e6af589e" x_b3_parentspanid:"-"

© 2020 General Electric Company 35

Page 40: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:40:16.968+0000] "GET /api/v1/analytic/execution/async/04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0/result HTTP/1.1" 200 0 283 "-" "Apache-HttpClient/4.5.3 (Java/1.8.0_152)" "10.72.11.111:19752" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.11.111" x_forwarded_proto:"https" vcap_request_id:"e8fcb852-c80d-46bf-590f-063abb6a07e2" response_time:2.441150172 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"915b9755bdeadd06" x_b3_spanid:"915b9755bdeadd06" x_b3_parentspanid:"-"

429fc52c-f855-4b81-8d7c-ce50e68505c2.run.aws-usw02-pr.ice.predix.io - [2019-03-07T04:40:19.579+0000] "DELETE /api/v1/analytic/execution/async/04-138d53c1-4093-11e9-bc8b-46c4f5ad01ea-sid-62fbba9e-fd90-4a49-ac3e-41dfd16db78a-0 HTTP/1.1" 204 0 0 "-" "Apache-HttpClient/4.5.3 (Java/1.8.0_152)" "10.72.11.111:19752" "10.72.53.197:61743" x_forwarded_for:"34.208.124.39, 10.72.11.111" x_forwarded_proto:"https" vcap_request_id:"639a9ef4-4494-4279-60a7-9520b34fca6c" response_time:0.195160282 app_id:"1fbd8a6f-dbbe-4b39-b8d1-bf9a7fe7d98f" app_index:"0" tenant:"-" x_b3_traceid:"7229fc83a49ff7b8" x_b3_spanid:"7229fc83a49ff7b8" x_b3_parentspanid:"-"

Delete Supporting Files from the Analytic Template

You can delete a supporting file that is no longer needed from the analytic template.

Procedure

1. In the module navigation menu, go to Analytics > Analytics Templates.2. In the Search pane, locate and select an analytic template to access it.3. Select the Analytics Template tab.4. Hover over a file listed in the SUPPORTING FILES subsection, then select X. You must confirm or

cancel your choice in the window that appears.5. Select Save to save changes to the analytic template. Select Cancel to discard all changes. You must

confirm or cancel your choice in the window that appears.

Results

After a successful save operation, the supporting file is removed from the list.

Add Analytic Definitions to the Analytic Template

Use the analytic template to define inputs, constants, and outputs found within the analytic runtime.

About This Task

This task does not apply if the template was ingested through the Import from file option on the NewAnalytic page. The configuration in the ingestion .json determines whether or not the configurations canbe changed via the UI.

36 © 2020 General Electric Company

Page 41: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

You will need to provide definitions for inputs and outputs in the analytic template. You can also find theconstants and their values in the analytic. The analytic template expects outputs events if your analyticmay produce alerts. At least one input and one output are required to proceed.

If you are adding inputs individually, make sure you indicate if the input is required. This information isused during I/O Mapping for Orchestrations.

Note: If you choose a SmartSignal runtime when creating an analytic, the Required column is read-only.

If you are using .csv files to add definitions, the system does not permit the file to be loaded if the fileformat or data type is invalid. In this case, a window displaying the invalid entries appears, allowing you tofix the indicated errors and retry. An input required indicator is needed for .csv files.

Important: Once the deployment has been completed, the output tag alias cannot be edited.

Procedure

• To manually enter definitions for the inputs, constants, and outputs:

1. In Analytics Templates section, select an analytic in the catalog.

2. Select in the Input Definition, Constants, and Output Definition subsections to enter valuesone at a time.

3. In the Output Events subsection, hover over the Name box and enter an alert template name toadd a new output event.

Important: The output name is case-sensitive and must match an existing alert template name. Ifthe alert template does not exist, the analytic fails to run.

• To upload .csv files for definitions for the inputs, constants and outputs:

1. In the Analytic Template section, select in the Input Definition, Constants, and OutputDefinition subsections to download a template used to create the definitions file for bulk upload.

After you have created .csv files, select in each subsection, then select the Upload a CSV box toupload format definitions for each type one at a time.

Tip:

◦ Each input definition, output definition, and constant name must match those defined in thedeveloped analytic.

◦ The data types for input definitions, output definitions, and constants must match the expecteddata. Definitions use INTEGER, BOOLEAN, STRING, and DOUBLE data types. For example, whendefining a tag_temperature input tag that produces data as doubles, you must define theDataType as Double.

◦ If you choose the Edge runtime and if incoming data types are not the same as the data typesrequired by an Edge analytic, automatic data type conversion will be attempted.

◦ In the output definition, tag names must match the tag names expected in the time series data.2. In the Output Events subsection, select the Name box and enter an alert template name to add a

new output event.

Important: The output name is case-sensitive and must match an existing alert template name. Ifthe alert template does not exist, the analytic fails to run.

Analytic Definition Files

You can create .csv files containing the relevant inputs, constants, and outputs for your analytic instead ofentering them manually in the analytic template. To do this, you must save a copy of the .csv template andadd the appropriate definitions before uploading it to the analytic template. If you upload a file with thewrong format or incorrect values, a dialog box displaying the invalid entries appears, and the file is notloaded. You can fix the indicated errors, and retry.

© 2020 General Electric Company 37

Page 42: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Tip: You can use the apm-analytic-csv-generator found on GitHub if you have access. You canrequest access from your support team. The .csv generator parses a sample .json analytic input file andwrites out input and constant .csv files.

Related Information

https://github.build.ge.com/212348840/apm-analytic-csv-generator

Define Analytic Input CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple inputs through .csv upload. You can skip thisprocedure if you want to add inputs directly through the UI.

Configure the template to specify input tags for the analytic you are adding or have added to the analyticcatalog. This information is needed to run the analytic.

Procedure

1. In the Analytic Template section, select in the Input Definitionsubsection to download a copy ofthe .csv template and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the AnalyticInputDefTemplate.csv file as AnalyticInputDef_data_wind_trend.csv and save it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values.

Table 4: Input Definition Guidelines

This definition Uses this guideline

Name A valid tag name that matches the analytic input parameter. Itmust match the regular expression pattern ^[A-Za-z0-9-_]+$.

Description The tag description. If there is no description for the tag,define blank entries as empty table cells, or offset them in acomma-delimited set (, ,).

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING

Note: You must select string data type for data of typestring, character, enum, or grid.

• INTEGER• DOUBLE• BOOLEAN

38 © 2020 General Electric Company

Page 43: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

This definition Uses this guideline

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the tags and attributes of allassets and their children starting at the defined Entity Type.For example, if you configure the Entity Type for your analyticinput at the Site level, then all tags and attributes of the Siteand its children become available for mapping duringdeployment.

Required The value must be checked or unchecked. Thisindicates whether the input is required.

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.Define Analytic Constants CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple constants through .csv upload. You can skip thisprocedure if you want to add constants on the UI.

Configure the template to specify constants for the analytics you are adding or have added to the analyticcatalog. This information is needed to run the analytics.

Procedure

1. In the Analytic Template section, select in the Constantssubsection to download a copy ofthe .csv template and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the AnalyticConstantsTemplate.csv file asAnalyticConstantsTemplate_data_wind_trend.csv and save it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values.

Table 5: Constants Guidelines

This definition Uses this guideline

Name A valid tag name must match the analytic constantparameter. It must match the regular expression pattern^[A-Za-z0-9-_]+$.

Description The tag description or null. If there is no description for thetag, define blank entries as empty table cells, or offset them ina comma-delimited set (, ,).

© 2020 General Electric Company 39

Page 44: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

This definition Uses this guideline

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING• INTEGER• DOUBLE• BOOLEAN

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the attributes of all assetsand their children starting at the defined Entity Type. Forexample, if you configure the Entity Type for your analyticconstant at the Site level, then all attributes of the Site and itschildren become available for mapping during deployment.

Value (optional) If there is no default value, you must define blank entries asempty table cells or offset them in a comma-delimited set(, , ).

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.Define Analytic Output CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple outputs through .csv upload. You can skip thisprocedure if you want to add outputs on the UI.

Configure the template to specify output tags for the analytic you are adding or have added to theanalytic catalog. This information is needed to run the analytic.

Procedure

1. In the Analytic Template section, select in the Outputssubsection to download a copy of the .csvtemplate and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the output-def.csv file as output-def_data_wind_trend.csv andsave it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values. Make sure your outputs match your analyticdefinition.

40 © 2020 General Electric Company

Page 45: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 6: Outputs Definition Guidelines

This definition Uses this guideline

Name A valid tag name that matches the analytic constantparameter. It must match the regular expression pattern^[A-Za-z0-9-_]+$.

Description The tag description or null. If there is no description for thetag, define blank entries as empty table cells, or offset them ina comma-delimited set (, ,).

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING• INTEGER• DOUBLE• BOOLEAN

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the tags and attributes of allassets and their children starting at the defined Entity Type.For example, if you configure the Entity Type for your analyticoutput at the Segment level, then all tags and attributes ofthe Segments and its children become available for mappingduring deployment.

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.

© 2020 General Electric Company 41

Page 46: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

6Analytic OrchestrationTopics:

• Manage AnalyticOrchestrations

• Manage Analytic Data Flows

42 © 2020 General Electric Company

Page 47: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Analytic Orchestrations

About OrchestrationsUse the Analytics Orchestration module to define the order in which algorithmic operations areperformed on sensor data.

Once defined, you can add an orchestration with your analytic templates. The Orchestration Engine uses aset of analytic sequences or defined analytical method flows called orchestrations in which analytics areordered as part of the analysis. You can create multiple orchestrations that execute in a linear fashion.

Tip: If you have scheduled several orchestrations to execute and one of the orchestrations fails,Application Analytics will bypass the failed orchestration and initiate execution of the next orchestration.

You can create an orchestration and link to an asset context using asset filters during deployments. Theasset filters allow you to automatically view only those assets when performing I/O mapping.

You can select and modify an orchestration from the Orchestrations queue. Orchestrations include ananalytics data flow that can connect and configure multiple analytics so they can be deployed together ina sequence. The data flow within orchestration is defined by mapping data field inputs and outputs foreach analytic in the set using I/O Mapping.

At this time, users can construct orchestrations containing only analytics as steps in the Analytic DataFlow, and orchestrations cannot be nested within each other.

You can sort the queue by the date created or the last update date.

From the orchestrations queue, each orchestration displays the following information when available:

• Orchestration name• Owner• Date created or last updated

The Search capability allows you to filter the orchestrations queue by name.

The orchestrations detail page uses the following sections to provide functionality or information about aselected orchestration:

Orchestration headingContains basic information about the orchestration:

• Orchestration name• Date and time the orchestration was saved, or the last date and time the orchestration was

updated• Orchestration owner• Status (Deployed or Not Deployed)

Orchestration sectionIncludes the orchestration description and analytics data flow.

Deployments sectionIncludes a listing of the current deployment, deployment date, and deployment type. You can add anew deployment and update an existing deployment in this section.

© 2020 General Electric Company 43

Page 48: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Add Orchestrations

Use the Orchestration page to add and modify orchestrations.

Before You Begin

You must add required analytic templates to the catalog before using them in orchestrations. Also addasset filters before using them in orchestrations.

Tip: When creating an analytic that triggers alerts, you need specific permissions to create or access analert.

About This Task

Use orchestrations to create a sequence of analytics. After you create a new orchestration and use theorchestration in a deployment, you must add an asset filter to ensure that only the assets for theorchestration appear when performing I/O mapping.

Procedure

1. In the module navigation menu, go to Analytics > Orchestrations.

2. Select to access the New Orchestration workspace.3. Enter a unique name for the orchestration in the Name box.4. The user name appears as the default value in the Owner box. Enter a different name if desired.5. Enter a brief description of the orchestration in the Description box.6. Select a Runtime Type in the box.7. Select Submit.

The orchestration is added and the workspace for the newly added orchestration appears.

Next Steps

Add an Analytics Data Flow in the Orchestration section.Delete an Orchestration

Deleting an orchestration deletes the orchestration and its deployments.

About This Task

Warning: Deleting an orchestration cannot be undone.

Procedure

1. In the module navigation menu, go to Analytics > Orchestrations.2. In the Search pane, select an existing orchestration to access its workspace.

3. In the workspace, select in the page heading.4. In the Delete Orchestration window, select Delete to confirm.

Results

The orchestration is removed from the list after successful deletion.

44 © 2020 General Electric Company

Page 49: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Analytic Data Flows

About Analytic Data FlowsSelect and connect analytics from the analytics catalog to build analytic data flows.

Analytics Data Flows allow you to select analytics from the analytics catalog and place them into aspecified order to run sequentially. All configuration information for the selected analytic is retrieved fromthe analytic template in the analytic catalog.

Each analytic added is considered a step in the analytic data flow. The inputs, constants, and outputs forthe analytic are mapped on the I/O Mapping page. Each step in the data flow can contain multipleiterations of the step's I/O mapping.

When an analytic is created and added to the analytic catalog, a version is assigned to that analytic. Forexisting deployments which contain analytics whose version has been updated (single-step and/or withinan Orchestration), and if the version has been marked as current, the deployment will then use theupdated analytic version. The Analytics Data Flow section lists the version of the analytic being used fora particular analytic step in an orchestration.

When you edit or remove an analytic from the Analytic Data Flow, existing deployments for theorchestration are not affected. An existing deployment reflects the change only after you redeploy theanalytic.

Note: If you do not wish to update an existing deployment with recent changes to analytics, then,consider creating an additional orchestration rather than redeploying an existing orchestration.

Determine Analytic Versions

You can determine the analytic version being used within an Analytic Data Flow.

About This Task

Determining an analytic version for an analytic step in an Analytic Data Flow allows you to makedecisions about modifying the orchestration based on that analytic version.

The analytic version refers to the analytic artifacts that were loaded into the analytic catalog. You candetermine the version being used by analytic orchestration, and the user who loaded it into the analyticcatalog. You can change the analytic version in an analytic orchestration data flow during deployment.

Procedure

• To determine the current analytic version in an analytic orchestration, do the following:

1. In the module navigation menu, go to Analytics > Orchestrations.2. In the Search pane, select an existing orchestration to access its workspace.3. In the Orchestration section, hover over a step in the Analytic Data Flow subsection.

The version for the analytic used in the step appears above the step.• To determine the analytic version of available analytic templates for use in orchestrations, do the

following:

1. In the module navigation menu, go to Analytics > Orchestrations.2. In the Search pane, select an existing orchestration to access its workspace.

3. Select the Edit icon ( ) in the Analytic Data Flow subsection of the Orchestration section.In the pane that appears, available analytic templates are listed with their version numbersbeneath each analytic.

© 2020 General Electric Company 45

Page 50: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Add an Analytic to a Data Flow

You can add one or more analytics in the catalog to the data flow.

Before You Begin

Make sure at least one analytic exists in the analytic catalog.

About This Task

You can add multiple analytics to the data flow. When you add or drag an analytic to an existing data flow,the newly placed analytic links automatically to the adjacent analytics. Each added analytic is considereda step in a sequence. You cannot reorganize the items to modify the sequence of the data flow. However,you can delete analytics in the data flow and add them back in a different order.

Note: Analytics added in the available analytics list also contain template information added in theanalytic catalog.

Warning: Adding, deleting, or changing the order of steps affects the current deployments.Consider creating a new orchestration rather than modifying an existing one.

Procedure

1. In the module navigation menu, go to Analytics > Orchestrations.2. In the Search pane, select an existing orchestration to access its workspace.3. Do one of the following to add the analytics template to the data flow:

Option

To add an analytic to an empty analytic data flow• Select the Add icon ( ) in the Analytic Data Flow

subsection of the Orchestration section.• In the pane that appears, locate the analytic that you

want to add as the first step in the analytic data flow.

Select the Add icon ( ) next to the analytic to insert, ordrag the new analytic to the Analytics Data Flowsection.

To add an analytic next to an existing analytic step in thedata flow • Select the Edit icon ( ) in the Analytic Data Flow

subsection of the Orchestration section.• Select the analytic in the data flow after which you want

to insert the new step.• In the pane, locate the analytic that you want to add as

the next step in the analytic data flow. Select the Add

icon ( ) next to the analytic to insert, or drag the newanalytic below the highlighted analytic in the AnalyticsData Flow section.

To add an analytic as the last step in the data flow• Select the Edit icon ( ) in the Analytic Data Flow

subsection of the Orchestration section.• In the pane that appears, locate the analytic that you

want to add as the last step in the analytic data flow.

Select the Add icon ( ) next to the analytic to insert, ordrag the new analytic to the Analytics Data Flowsection.

46 © 2020 General Electric Company

Page 51: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Modify an Analytic Data Flow

You can modify the analytic data flow to add existing analytic template.

About This Task

When you modify an analytics data flow, make sure that the orchestration is not deployed. Considercreating a new orchestration rather than modifying one that is in use.

In the Analytic Data Flow edit page, you can:

• Modify the description.• Add analytics in the Available Analytics list.• Delete analytics in the Analytic Data Flow.

Make sure you add analytics in the order you want them to execute.

Procedure

1. In the module navigation menu, go to Analytics > Orchestrations.

2. In the Search pane, select an existing orchestration to access its workspace.

3. Hover over the existing description, then select on the Description box, then make yourmodifications, and then press ENTER or select outside the Description box.

4. Select the Edit icon ( ) in the Analytic Data Flow subsection of the Orchestration section.

5. Add analytics or delete analytics in the data flow.

6. Select Save to save your modifications.

Delete an Analytic in a Data Flow

You can delete an analytic step in an analytic data flow.

About This Task

Warning: Adding, deleting, or modifying the order of steps affects the current deployments.Consider creating a new orchestration rather than modifying an existing one.

Procedure

1. In the module navigation menu, go to Analytics > Orchestrations.

2. In the Search pane, select an existing orchestration to access its workspace.

3. Select the Edit icon ( ) in the Analytic Data Flow subsection of the Orchestration section.

4. In the Analytics Data Flow section, hover over an analytic step, and then select .

5. In the Confirmation dialog box, select Remove.

6. Select Done in the page header to exit the edit mode.The analytic step is removed in the data flow. The remaining analytic steps in the data flowautomatically reconnect.

Modify the Analytic Version During Deployment

You can modify the analytic version being used within an Analytic Data Flow during deployment.

Before You Begin

You must have uploaded more than one analytic version to modify during the deployment of that analytic.

© 2020 General Electric Company 47

Page 52: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About This Task

Modifying the analytic version for an analytic step in an analytic data flow allows you to use a differentversion than set as current (default) in the analytic template.

The analytic version refers to the analytic artifacts that were loaded into the analytic catalog. You candetermine the version that the analytic uses when deployed by itself or within an orchestration.

Tip: Map your inputs, constants, outputs, and attributes after you have modified the version of theanalytic to apply during deployment. Modifying a different version post mapping may require you toperform the mapping step one more time for properly deploying it to the runtime.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the deployment.4. Select the 2. I/O Mapping tab.

5. Select to access the Data Flow pane.6. Select a analytic step node. For example, if you have Analytic_Wind_Throughput as one of the analytic

step in the deployment, then select that node to access Analytic_Wind_Throughput subsection of the2. I/O Mapping section in the deployment.

7. Select a version in the drop-down list box in the Version column, and then select Use this Version.The Mapping may be lost box appears with the message If there are structuralchanges between analytic versions, you may lose some or all of theexisting mapping.

8. In the Mapping may be lost box, select Continue to change the version or select Cancel to keep thecurrent version.

48 © 2020 General Electric Company

Page 53: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

7Analytic DeploymentsTopics:

• Manage Analytic Deployments• Select Asset Applicability for

Analytics• Map Analytics Inputs and

Outputs to Tags• Manage Analytic Iterations

within an Orchestration• Manage Deployment Schedules• Review Deployments

© 2020 General Electric Company 49

Page 54: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Analytic Deployments

About Analytics DeploymentsYou can create, configure, schedule, and execute analytic deployments that will be used to trigger eventsused by Alerts.

A successful deployment pushes the deployment configuration files to the cloud, and the analytic is run asscheduled or on-demand.

An analytic deployment requires the following general steps. Complete each step before proceeding.

Asset SelectionSelect the target assets for the analytic deployment.

I/O MappingMap inputs and outputs in your analytic template definitions to ingested asset tags. Add constants inthe selected analytic runtime. Add an alert template name that will use the analytic to trigger alerts.

SchedulingSchedule ongoing analytic runs, or run an analytic on-demand.

ReviewAccess, modify, and validate your selections.

Add a Deployment

You can add a deployment configuration to deployments prior to configuring and scheduling.

Before You Begin

• Make sure you have ingested assets into the system with all the required tags and attributes neededfor I/O mapping.

• For each analytic being deployed to a set of assets, make sure you have added the template to thecatalog, and that you have defined the analytic inputs, constants, outputs, and output events asneeded to run the analytic.

Note: Not all analytic deployments will need constants and output events.

Procedure

1. Select one of the following ways to add a deployment.

• In the Deployments page:

a) In the module navigation menu, go to Analytics > Deployments.

b) In the heading, select .

• In the Orchestrations page:

a) In the module navigation menu, go to Analytics > Orchestrations.b) In the Search pane, select an orchestration to access its workspace.c) In the workspace of the selected orchestration, select the Deployments tab.

d) In the DEPLOYMENTS subsection, select .

• In the Analytics Templates page:

50 © 2020 General Electric Company

Page 55: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

a) In the module navigation menu, go to Analytics > Analytics Templates.b) In the Search pane, select an analytic template to access its workspace.c) In the workspace of the selected analytic template, select the Deployments tab.

d) In the DEPLOYMENTS subsection, select .

2. In the New Deployment window, select one of the following:

• Analytic Template• Orchestration

These choices are not available when adding a deployment directly to a selected analytic template ororchestration.

3. Enter a unique deployment name in the Deployment Name box.

Note: Make sure you enter a unique name. The system checks for duplicates.

4. Enter the name that you want to use for the deployment in the Template Name box.

Note: This field will be pre-populated with the template or orchestration name depending on whereyou initiate the deployment creation. If pre-populated, you cannot modify the value.

5. Optional: To apply a different version of the analytic template than the default version, select aversion in the Template Version drop-down list box.

Note: This option is only available for deployments of the analytic template, not for orchestrations.The box is disabled if only one version of the template artifact is present in the catalog.

Search and Filter Deployments

Perform this task to search for related deployments by asset ID or filter by deployment or analytictemplate name.

Before You Begin

You have access to analytic deployments. You have at least one available deployment.

About This Task

If you want to narrow the list of deployments or find deployments that are applicable to a specific asset,you can use the search and filter option available on the Deployments page. You must have a valid AssetID before using it in the search. The Asset ID is usually the asset instance or entity's source key. If theAsset ID belongs to an asset that has one or more attributes that were used to define an asset filter, thenall deployments that uses that asset filter appear as the query result.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. Enter an Asset ID in the Enter an Asset ID box.

The Asset ID entered must exist in the system and must match exactly. Use of partial keywords or useof wild cards in the search are not supported.

Results are provided with deployments that uses that asset, or deployments that uses an asset filterthat has one or more attributes as that asset. If the asset does not exist then the message AssetNot Found appears.

3. Optional: Enter characters in the Search deployments box to narrow the number of deploymentsdisplayed.

Results are provided as you enter characters. Result matches for the search text contained within theName or Template columns appear. For example, if you searched for the text turbine, then

© 2020 General Electric Company 51

Page 56: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

deployments that have the Analytics Template name or Deployment name containing turbine willappear.

4. Optional: Perform any of the following optional steps:To do this Do this

Change the number of results per page Select a value in the Rows per page box to display more orless results. The maximum number of results per page is 20.

Navigate between pages of results If the number of results exceed the maximum page display,the results are distributed across multiple pages. Use thepage number link or the arrow links to navigate betweenpages.

Delete a Deployment

You can delete a deployment you created if it is no longer needed. Consider changing the deploymentschedule instead of deleting it. Deleting a deployment cannot be undone.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the workspace of the selected deployment, select the Delete button.4. In the Delete Deployment window, select Force Delete, and then select Submit to delete the

deployment and its configuration.

Deployment Statuses

List of possible deployment configuration and job deployment statuses.

Application Analytics maintains the status of the analytic deployment staging in the runtime engine. Ifstaging in the analytic runtime fails, you will see Failed in the Status column on the Deployments page.You can also see the last time the deployment was saved in the Last Updated column.

The following possible outcomes to the staging process can be viewed in the Deployments page:

Not DeployedThe deployment was created but is not ready to be staged in the configured runtime for execution.

In ProgressAnalytic artifact you uploaded is staged in the configured analytic runtime ready for running thedeployment.

FailedA back-end issue such as a network interruption has occurred. Return to the Review section and tryagain.

Run OnceThe deployment is validated and pushed to the cloud, where it is running. In the Search pane of theAnalytics Templates page, you will see the date the deployment was last updated next to theanalytic name.

ScheduledAnalytic artifact you uploaded is staged in the configured analytic runtime ready for running at thescheduled interval as configured in the deployment.

Note: In the Deployments section of the Analytics Templates page, status messages are displayednext to the deployment name in the Current Deployments subsection.

Related InformationCannot Deploy an Analytic to Predix Runtime on page 77

52 © 2020 General Electric Company

Page 57: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

You cannot deploy an analytic even after completing all the required steps in deployment.

Select Asset Applicability for AnalyticsIn this section, you will find information on selecting the asset applicability for the analytic deployment.

Select Assets for DeploymentSelecting assets for deployment configures asset applicability for your deployment.

Before You Begin

• Use the Asset Filters link on the navigation menu to create asset filters that can be applied on theAsset Selection tab.

• Make sure you have added and named the deployment.

About This Task

During asset applicability configuration, GE Digital APM allows you to:

• Select assets from a list of available equipment templates. This option is only available for tenants thatare configured to support selection through equipment templates.

• Select from all the assets available in the enterprise.• Select the assets within an asset filter.• Select a subset of the assets within an asset filter.• Search for asset attributes to modify asset filter sets.• Create a new asset filter.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select a deployment in the list to access it.3. Choose a method for selecting assets for deployment.4. Select Next or the I/O Mapping tab.

Results

The assets selected for deployment are saved.

Next Steps

Do one or more of the following tasks:

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61• Map an Attribute on page 65

Apply an Asset Filter to the Deployment

Use an Asset Filter to apply assets to your deployment.

Before You Begin

Make sure you have added and named the deployment.

© 2020 General Electric Company 53

Page 58: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About This Task

Use the predefined filters in the Asset Filters box to select the asset applicability for the deployment.You can search for asset attributes that will further narrow your asset filter selection.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the Asset Filter box, and then select an asset filter in the list.5. Optional: Enter characters in the Search box.

You can enter multiple search criteria to narrow your search results. Each key-value pair represents asingle search criterion. You can select the X on the search text to remove it from the list.

Suggestions are provided as you enter characters. Select an attribute from the suggestions. Enter anequal to (=) operator and a value to search (for example, Asset Type=21000) and then press Enter toadd the search criterion.

6. Select Search.

If multiple search keywords are used in the Search box, then the and operator is used to filter results.Only results matching all search criteria are listed. Remove search criteria to widen the search.

Search results are a subset of the already applied asset filter that you previously selected.7. Select Next or the 2. I/O Mapping tab.

ResultsThe assets selected for deployment are saved.

Next StepsDo one or more of the following tasks:

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61• Map an Attribute on page 65

Apply Assets from Equipment Template

You can select assets from a list of available equipment templates.

Before You Begin

This procedure assumes that the following prerequisites tasks have been completed.

• You have uploaded the required assets, tags, equipment templates, and time series data required forthis analytic.

• Your tenant has been configured to support asset selection through equipment templates.• You have the required access permissions to manage analytics on the selected runtime for your

tenant.• You have added and configured the analytic template and/or orchestration in the analytics catalog.• You have created a deployment.

About This Task

Selecting the equipment template for deployment allows you to access tags and attributes for thatequipment template during the I/O Mapping step in deployment. Such deployments will be created foreach asset instance within the selected and filtered equipment template.

54 © 2020 General Electric Company

Page 59: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select a value in the Equipment Template box.5. Optional: Select a value in the Asset Filter box to further filter the list.

Note: The selected filter must be of the same asset classification type as configured within theselected equipment template to produce results. If the selected filter and equipment template havemismatched asset types, no results will appear.

A table of the available assets included in the selected equipment template appears. All assets in theresult set are selected for deployment. During runtime, analytics will run on all assets selected in thisprocedure. You cannot manually remove the applicability but can filter further to narrow your resultset.

6. Optional: After selecting an asset filter, select Search.

Note: In this step during deployment, the Search box is disabled. You cannot further filter resultsusing the keyword search after selecting an equipment template.

7. Select Next or the 2. I/O Mapping tab.

Results

The assets selected for deployment are saved.

Next Steps

Do one or more of the following tasks:

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61• Map an Attribute on page 65

Apply a Subset of an Asset Filter to a Deployment

You can select a subset of an Asset Filter to apply to a deployment.

Before You Begin

Make sure you have added and named the deployment.

About This Task

Instead of creating a new Asset Filter for deploying assets, you can use an existing Asset Filter to selectonly the assets needed for the deployment. You can search for asset attributes that will further narrowyour asset filter selection.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the Asset Filter box, and then select an asset filter in the list.5. Optional: Enter characters in the Search box.

You can enter multiple search criteria to narrow your search results. Each key-value pair represents asingle search criteria. You can select the X on the search text to remove it from the list.

© 2020 General Electric Company 55

Page 60: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Suggestions are provided as you enter characters. Select an attribute from the suggestions. Enter anequal to (=) operator and a value to search (for example, Asset Type=21000) and then press Enter toadd the search criterion.

6. Select Search.

If multiple search keywords are used in the Search box, then the and operator is used to filter results.Only results matching all search criteria are listed. Remove search criteria to widen the search.

Search results are a subset of the already applied asset filter that you previously selected.7. Enter characters in the Search box to find assets.

Assets results are provided as you enter characters.8. Perform any of the following optional steps:

To do this Do this

Change the number of asset results per page Select a value in the Rows per page box to display more orless results. The maximum number of results per page is 20.

Navigate between pages of the asset results If the number of results exceed the maximum page display,the results are distributed across multiple pages. Use thepage number link or the arrow links to navigate betweenpages.

9. Do one of the following:

• Clear the Select All box before the Name column to deselect all assets in the filter, then select thecheck boxes for the assets you want to include. As you make choices, the number of selectionsappear above the table. You can choose up to twenty assets before proceeding to the next step.

• Clear the check boxes for assets you want to exclude. As you clear the check boxes, the number ofselections appearing above the table decreases.

Note: You can search for assets by entering at least three characters into the Search box. Assetsresults are provided as you enter characters.

10. Select Next or the 2. I/O Mapping tab.

Results

The assets selected for deployment are saved.

Next Steps

Do one or more of the following tasks:

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61• Map an Attribute on page 65

Apply Assets from the Enterprise to a Deployment

You can select assets from all the available assets in an enterprise to apply to a deployment.

About This Task

Use the Asset Filter to apply only the assets needed within the enterprise for the deployment. You cansearch for asset attributes that will further narrow your asset filter selection.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.

56 © 2020 General Electric Company

Page 61: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

4. Select the 1. Asset Selection tab.

5. Select the Asset Filter box and select All Assets in the list.A table of the available assets included in the enterprise appears.

6. Optional: Enter characters in the Search box to find asset attributes specific to the selected assetfilter, then select Search.

Suggestions are provided as you enter characters.

7. Do one of the following:

• Clear the check box before the Name column to clear all selection, then select the check boxes forthe assets you want to include. As you make choices, the number of selections appear above thetable. You can choose up to 200 assets before proceeding to the next step.

• Clear the check boxes for assets you want to exclude. As you clear the check boxes, the number ofselections appearing above the table decreases.

Note: You can search for assets by entering at least three characters into the Search box. Assetsresults are provided as you enter characters.

8. Perform any of the following optional steps:To do this Do this

Access only the selected assets in the table Select the <number> Selected link above the table. To returnto the full table, select Add Assets.

Access all the available assets for the asset filter Select the Display Assets link above the table.

Collapse the asset selection table Select the Hide All link above the table.

9. Select Next or the 2. I/O Mapping tab.

Results

The assets selected for deployment are saved.

Next Steps

Do one or more of the following tasks:

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61• Map an Attribute on page 65

Map Analytics Inputs and Outputs to Tags

About I/O MappingUse I/O Mapping to connect, or map, the applicable asset tags in the analytics deployments for analyticsinputs, outputs, and constant definitions that were previously configured in the Analytic Template.

Input tags are marked as optional or required in the 2. I/O Mapping section. You cannot proceed to thenext step until all inputs marked as required are mapped.

You can only add or delete orchestration inputs through the 2. I/O Mapping section if your AnalyticTemplate configuration has this enabled. If not, you must modify the original analytics definitions in theAnalytic Template.

Important: The changes in the analytics catalog and template affect all analytics deployments andorchestrations that are using the selected analytic. Consider creating a new analytic rather thanmodifying one that is in use.

© 2020 General Electric Company 57

Page 62: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Map Tags and Asset Attributes for Deployment

The I/O Mapping is the second step in deployment. It allows you to map the inputs that were defined whilebuilding the analytic template to tenant-specific asset tags and attributes.

Mapping an analytic to time series data or asset attribute involves associating an analytic input to assettags or assigning values including attributes.

You must select at least one asset, or an asset filter in the 1. Asset Selection section of deployment.

Tip: Input, constant, and output definitions are created while building the analytic template. Any changesmade while building the analytic template are reflected in the related sections.

The framework does not allow partial mapping for required inputs. All required inputs must be mapped foreach selected asset before you can proceed. If you are mapping more than 20 assets, it is recommendedto map inputs by common tags. Any template definitions for constants and outputs are automaticallymapped; however, you can replace specific values.

The 2.I/O Mapping section contains the following components:

Component Description

ASSET(S) pane The ASSET(S) pane may be hidden. To access this pane, select

to expand it or select to collapse it. Depending onthe type of mapping, you can access either the common tagsavailable for all selected assets in the deployment or the assethierarchy with only the relevant tags within that single assetnode.

• ASSET(S) subsection - Displays the asset applicability.Allows you change the asset selection to access tags andattributes of a specific asset.

• Data Sources box- Allows you to select from available datasources. This selection varies based on the deployment.

• Search expanded tags box - Allows you to search for tagswithin the context of the expanded Tags node. For example,if you have expanded the Tags node for asset K12Turbine 258, then the context for search is limitedto tags within that asset.

• Common subsection - Accessible only when All Assetsoption is selected. Displays the asset tree with Tags,Custom Attributes, and Reserved Attributes nodes. Onlythe common tags and attributes are available for mapping.

• Find by Name subsection - Accessible only when All Assetsoption is selected. Allows you to search for specific tags orattributes by name.

• - Accessible only when a specific asset is selected.Displays the asset tree with Tags, Custom Attributes, andReserved Attributes nodes. You can access tags, customattributes, and reserved attributes for the selected assetonly.

Data Flow pane The Data Flow pane may be hidden. To access this pane, select

to expand it or select to collapse it. Adding iterationsis applicable to orchestration deployments.

INPUT DEFINITION: TAGS & ATTRIBUTES table A table of inputs to be used in the deployment. You can map therequired tags to inputs here. When you are mapping asset byasset, only inputs for the selected asset appear in the ASSETSsubsection.

58 © 2020 General Electric Company

Page 63: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Component Description

CONSTANTS table A table of constants selected when building the AnalyticTemplate. You can change the value for constants displayedhere.

OUTPUT DEFINITION: TAGS & ATTRIBUTES table A table of outputs selected when building the AnalyticTemplate. You can select the Output UoM and UoM Conversionvalues, map a tag to the constant, and select the outputsdestination in a drop-down list. If enabled, you can mapaggregate tags for computing aggregate values.

OUTPUT DEFINITION: ALERTS table A table of Alert Template associations selected when buildingthe Analytic Template.

Modify the Analytic Version During Deployment

You can modify the analytic version being used within an Analytic Data Flow during deployment.

Before You Begin

You must have uploaded more than one analytic version to modify during the deployment of that analytic.

About This Task

Modifying the analytic version for an analytic step in an analytic data flow allows you to use a differentversion than set as current (default) in the analytic template.

The analytic version refers to the analytic artifacts that were loaded into the analytic catalog. You candetermine the version that the analytic uses when deployed by itself or within an orchestration.

Tip: Map your inputs, constants, outputs, and attributes after you have modified the version of theanalytic to apply during deployment. Modifying a different version post mapping may require you toperform the mapping step one more time for properly deploying it to the runtime.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the deployment.4. Select the 2. I/O Mapping tab.

5. Select to access the Data Flow pane.6. Select a analytic step node. For example, if you have Analytic_Wind_Throughput as one of the analytic

step in the deployment, then select that node to access Analytic_Wind_Throughput subsection of the2. I/O Mapping section in the deployment.

7. Select a version in the drop-down list box in the Version column, and then select Use this Version.The Mapping may be lost box appears with the message If there are structuralchanges between analytic versions, you may lose some or all of theexisting mapping.

8. In the Mapping may be lost box, select Continue to change the version or select Cancel to keep thecurrent version.

© 2020 General Electric Company 59

Page 64: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manually Map Inputs for Selected Assets (All Assets)

Use the Mapping Options box to map inputs for tags common to all assets selected in first step inDeployment.

Before You Begin

Make sure you have selected the target assets in the first step in Deployment.

About This Task

When you select the option to map inputs for all assets, the tags common to all assets selected in the 1.Asset Selection section are available to map. Mapping for all assets will map the inputs for the assetsthat use the selected tag.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 2. I/O Mapping tab.

5. Select to access the ASSET(S) pane.6. Select the Change Selection button.7. In the list, select the value All Assets.

The value in the ASSET(S) pane displays the selected value.8. In the Common subsection, select the Tags node to access tags available for mapping.

• Navigate the tree structure to find specific tags to map.• Or enter characters in the Search expanded tags box to find tags. Results are provided as you

enter characters.

Note: If there are no tags common to the selected assets, you must map by asset.9. Select a tag, drag and drop it onto the input row to map it.10. Perform any of the following optional steps in the INPUT DEFINITION: TAGS & ATTRIBUTES table:

To do this Do this

Change the number of inputs per page Select a value in the Rows per page box to display more orless results. The maximum number of inputs per page is 20.

Navigate between pages of the inputs If the number of inputs exceed the maximum page display, theresults are distributed across multiple pages. Use the pagenumber link or the arrow links to navigate between pages.

11. Repeat mapping tags for each required input definition.

Important:

• The MAPPING KEY column indicates which inputs are required.• Map one input per tag. Do not map the same tag to multiple inputs for the selected asset.• You cannot proceed until all required tag definitions are mapped.

12. Select Save Changes to apply and save the mapping configuration.

60 © 2020 General Electric Company

Page 65: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manually Map Inputs per Asset (Asset by Asset)

Map inputs to tags associated with a single asset applied in asset selection during deployment.

Before You Begin

Make sure you have selected the target assets in the 1. Asset Selection section for the deployment.

About This Task

Use this procedure to map tags or attributes for each selected asset individually. After an asset is selectedfor mapping, only input definitions for that asset are displayed in the INPUT DEFINITION: TAGS &ATTRIBUTES table, and only the selected asset tags and asset attributes are available as nodes in the

Tags and Attributes pane. displays tags and attributes for the selected asset.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 2. I/O Mapping tab.

5. Select to access the ASSET(S) pane.6. Select the Change Selection button.7. Select an asset value in the list.

The value in the ASSET(S) pane displays the selected value.

8. In , select the Tags node to access tags available for mapping.

• Navigate the tree structure to find specific tags to map.• Or enter characters in the Search expanded tags box to find tags. Results are provided as you

enter characters.9. Select a tag, drag and drop it onto the input row to map it.10. Perform any of the following optional steps in the INPUT DEFINITION: TAGS & ATTRIBUTES table:

To do this Do this

Change the number of inputs per page Select a value in the Rows per page box to display more orless results. The maximum number of inputs per page is 20.

Navigate between pages of the inputs If the number of inputs exceed the maximum page display, theresults are distributed across multiple pages. Use the pagenumber link or the arrow links to navigate between pages.

11. Repeat mapping tags for each required input definition, then select the Change Selection button inthe ASSET(S) box to select another asset.

Important:

• The Tags column indicates which inputs are required.• Map one input per tag. Do not map the same tag to multiple inputs for the selected asset.• You cannot proceed until all required tag definitions are mapped.

12. Select Save Changes to apply and save the mapping configuration.

© 2020 General Electric Company 61

Page 66: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Map Inputs with Suggested Tags

Use the Suggested Tags option to map inputs for tags that have the same name as the input.

Before You Begin

Ensure that you have selected the target assets in the first step in deployment.

About This Task

Automapping maps the input and constant ports of an analytic with the tags or attributes of an asset thatmatches the names of the ports. Depending on the Mapping Value Type (MVT), that is, a tag or attribute ofeach input and constant port of an analytic, automapping searches for the corresponding tag or attributethat matches the port name in the assets that are a part of the asset filter, and then assigns the tag orattribute to each port. In addition, the data source corresponding to a tag or attribute is automaticallyselected for each input, constant, and output ports.

In addition to APM Time Series and APM Asset data sources, any data sources that are used must beregistered with TMS. CAF recognizes two types of tags, which are Sensor and Computed. Sensor tags areassigned to input ports, and Computed tags are assigned to output ports. The tenant must assign adefault data source for each type of tags and attributes.

If you have configured input parameters similar name as tags available for mapping, then you may use theSuggested Tags option to speed up the mapping process.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 2. I/O Mapping tab.5. In the INPUTS: TAGS & ATTRIBUTES table, select the Get Suggested Mapping button to access tag

suggestions to map to inputs.6. If you accept the suggestions, select Save Selected. If you do not want to accept suggestions or if no

suggestions are listed, select Reset.7. Repeat for each required input tag.

Important:

• The Tags column indicates which inputs are required.• Map one input per tag. Do not map the same tag to multiple inputs for the selected asset.• You cannot proceed until all required tag definitions are mapped.

8. Select Save to apply and save the mapping configuration.

Map Multiple Tags to an Input

You can map a single input port to multiple tags at any level of the asset hierarchy.

Before You Begin

This procedure assumes that the following prerequisite tasks have been completed.

• You have uploaded the required assets, tags and time series data required for this analytic.• You have configured asset filters to use in the deployment.• You have the required access permissions to manage analytics on Spark runtime.• You have added and configured an Analytic Template that has the Aggregate Mapping option enabled.

62 © 2020 General Electric Company

Page 67: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

• You have created a deployment for that Analytic Template.• You have applied the deployment to the applicable assets through asset selection.

About This Task

This functionality is only available for tenants that are configured to support running analytics on theSpark runtime cluster environment. Depending on the tenant configuration, this option may not beavailable. The analytic artifact uploaded to the Analytic Template being used in the deployment also mustsupport aggregation.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. Select an existing deployment to access its workspace.

3. In the page heading, select Edit to access the 1. Asset Selection section.

4. Select the 2. I/O Mapping tab.

5. Select to access the ASSET(S) pane.

6. Select the Change Selection button.

7. In the list, select the value All Assets.The value in the ASSET(S) pane displays the selected value.

8. In the Common subsection, select the Tags node to access tags available for mapping.

• Navigate the tree structure to find specific tags to map. You can navigate at any level in the assethierarchy including tag nodes of child assets. For example, if the Site is the parent node in the assethierarchy, you can access common tags at its segment and asset level.

• Or enter characters in the Search expanded tags box to find tags. Results are provided as youenter characters.

Note: If there are no tags common to the selected assets, you must map by asset.

9. Select a tag, drag and drop it onto the input row to map it.

10. Select another tag, drag and drop it onto the same input row to map it too.

Results

The value in the MAPPING KEY column for that input changes to Multiple. Hovering your mouse cursorover the text shows a comma-separated list of tags that are currently mapped to that input.

Add Mapping Override to Analytic Inputs

You can add asset applicability overrides to inputs during I/O Mapping step in deployment.

Before You Begin

Make sure you have added at least one deployment. This is a tenant and runtime specific option. Makesure that the current runtime for the Analytic Template used in the deployment allows adding overrides.

About This Task

This option is only available for tenants that are configured to support input overrides during deployment.Using overrides allows you to select a different asset applicability than that defined in the 1. AssetSelection section. It is recommended that you map tags to inputs both for its default asset applicabilityand overriding asset applicability.

Inputs are marked as optional or required in the 2. I/O Mapping section of the Deployments page.

© 2020 General Electric Company 63

Page 68: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. Select an existing deployment to access its workspace.

3. In the page heading, select Edit to access the deployment.

4. Select the 2. I/O Mapping tab.

5. Select to expand the Data Flow pane.

6. Navigate to an iteration node to select it. For example, if you have Analytic_Wind_Throughput as one ofthe analytic steps in the deployment, then expand the node and select the iterationAnalytic_Wind_Throughput_Iteration_1 to access it.

7. In the INPUT DEFINITION: TAGS & ATTRIBUTES table, hover over the MAPPING KEY column of the

input row to which you wish to add overrides. Select the to access the OVERRIDES table.

8. Select the + Add Overrides button to add a new row.

9. Select the Select an Asset Filter button.

10. In the Override Criteria window that appears, select the check boxes of one or more filters listed, andthen select the Add button.The selected asset filters appear in the Asset/Filter column.

11. Use one of the following methods to map inputs to tags.

• Manually Map Inputs for Selected Assets (All Assets) on page 60• Manually Map Inputs per Asset (Asset by Asset) on page 61

Results

The icon appears in the MAPPING KEY column for that input. Hovering over shows a comma-separated list of overrides.

Map Inputs, Constants, and Outputs for Assets in an Equipment Template

Map inputs, constants, and outputs for all assets within the selected equipment template in step 1 for theAnalytic Deployment.

Before You Begin

This procedure applies to deployments using an equipment template for its asset applicability. Make sureyou have selected the target equipment template to derive assets to map in step 1 for the AnalyticDeployment.

About This Task

When you map inputs for assets within an equipment template, tags common to all assets selected instep 1 for the Analytic Deployment are available. If filters were applied to narrow the asset applicability,only tags common to those selected assets are available.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. In the Search pane, select an existing deployment to access its workspace.

3. In the page heading, select Edit to access the 1. Asset Selection section.

4. Select Next or the 2. I/O Mapping tab. You can map tags in one of the following ways:

64 © 2020 General Electric Company

Page 69: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Option Description

Mapping inputs, constants, and outputs to common tags The resulting list of tags are derived from an intersection of alltags within the template across all nodes. All common tagsacross the child nodes are listed at the root node. You can seeall common tags, or tags filtered by Asset Type (class), or tagsfiltered by Sub Templates. For example, if GE90-Template-AV is the root node (first position), Tagsnode at the root of the tree structure will list all common tags.

Mapping inputs, constants, and outputs to tags byposition

The resulting list of tags within a selected position.

a) Select to access the ASSET(S) pane.b) In the Common subsection, expand nodes within the template to reveal tags common to map. You

can traverse through the template structure to find common tags.

Note: Each node starting at the root have the following structure.

root (Position 1)|-tags|-attributes|-Sub Position 2 (equipment_1) |--tags |--sub template_1 |---sub position a (equipment_a) |---tags |---sub templates |--attributes |-Sub Position 2 (equipment2) |--tags |--sub template_2 |---sub position b (equipment_b) |---tags |---sub templates |--attributes

c) Select the tag. Drag and drop the tag on to the input definition, constant, or output definition rowto map the tag.

The Data Source row displays the name of the mapped tag.5. Repeat for each required input, constant, and output tag.

Note: You cannot proceed until all required tag definitions are mapped.6. Select Save to apply and save the mapping configuration.

Map an Attribute

You can map asset attribute to inputs, constants, and outputs in an analytic deployment.

Before You Begin

Make sure you have selected the target assets for your deployment.

The asset model has attributes ingested as part of the asset model.

© 2020 General Electric Company 65

Page 70: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About This Task

The data format for the component to be mapped (input, constant, or output) must be set to AssetAttribute when defining the analytic template. You cannot map attributes to data format defined as TimeSeries.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 2. I/O Mapping tab.

5. Select to access the ASSET(S) pane.6. Expand the @Reserved Attributes or @Custom Attributes nodes to select from a list of asset

attributes available.7. Select an attribute from the list, and then drag and drop it onto the component (input, constant, or

output) row to map.

Note: You will receive an error message if you drag and drop onto the row with Time Series dataformat. You can only map attributes to analytic definition with the Asset Attribute or ANY data format.

You must only select attributes that have the same data type as the defined. For example, if youranalytic expects a STRING value in its input Temp Threshold, the selected attribute should alsohave the data type of STRING.

The selected attribute is mapped in the Mapping Key column.8. Repeat this for mapping all other required definitions with Asset Attribute data format.

Modify the Output UoM Values

You can change a defined value for a unit of measure for the selected deployment. The unit ofmeasurement (UOM) values are defined when building the Analytic Template. Changing the unit ofmeasurements in the deployment affects the selected deployment only.

Before You Begin

Make sure you have selected the target assets for your deployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 2. I/O Mapping tab.5. In the OUTPUT DEFINITION: TAGS & ATTRIBUTES table, select the Output UoM box to access

output unit of measurements and then select a value in the list.6. Select Save to save the deployment.

66 © 2020 General Electric Company

Page 71: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Manage Analytic Iterations within an Orchestration

About IterationsUse iterations to apply the execution of the analytics algorithmic techniques multiple times (repeatedly)within the same step of the orchestration.

Iterations, or repeated executions, occur in parallel with each other, with each execution utilizing a specificset of inputs that the user will configure in the 2. I/O Mapping section in the deployment.

Iterations allow you to use I/O mapping to add mapping or copy a data flow step. You are able to:

• View the number of iterations and navigate between them.• Map inputs for selected assets to tags.• Edit constant values.• Map outputs for selected assets to tags.

You can view the number of iterations associated with an analytic step in the 2. I/O Mapping section ofdeployments. The Data Flow pane displays the analytic steps in the orchestration and its iterations in a

tree view. The Data Flow pane may be hidden. To access this pane, select to expand it or select to collapse it. Adding iterations is applicable to orchestration deployments.

The inputs and outputs for the iterations are mapped in the same way as the other I/O mappings. Someinputs, or outputs allow you to change the Unit of Measure. When mapping, you cannot use asset tagsthat do not match the Unit of Measure selected. Inputs, constant values, and selections for outputs can beedited for the iteration.

The asset tag and asset hierarchy tag list provides available asset tags based on the applicability definedin the orchestration.

Iterations apply only in the context of the selected orchestration deployment and step.

Add an Iteration to an Analytic Step

Add an iteration when you want to apply the execution of the analytic algorithmic techniques for theanalytic multiple times (repeatedly) within the same step of the orchestration, with each additionalexecution utilizing a specific set of inputs. In adding an iteration, you must map all the inputs and outputsspecifically for the new iteration.

About This Task

Input and output mappings are blank when you add an iteration. You must map all inputs and outputs forthe new iteration.

When you add an iteration to an existing orchestration deployment, previously deployed analytics are notaffected. You must redeploy the orchestration to change any existing deployments.

Note: Consider creating an additional deployment for the orchestration rather than modifying an existingdeployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. Select an existing deployment to access its workspace.3. In the page heading, select Edit to access the deployment.

© 2020 General Electric Company 67

Page 72: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

4. Select the 2. I/O Mapping tab.

5. Select to expand the DATA FLOW pane.

6. Select the button next to the analytic step node. For example, if you want to add an iteration to

Analytic_Wind_Throughput, select next to it.

7. Enter the iteration name in the box that appears, and then select to add it to the analytic stepnode.

8. Select the newly added iteration node to access its workspace.The 2. I/O Mapping section reflects the configuration for the selected iteration.

9. Enter values in the Data Points box to use for the iteration.

An analytic may require a discrete number of points for execution (for example, 30 points to properlycalculate the smoothed value of the particular metric). You can define custom values for data points touse at the iteration level.

Note: When running the orchestration, data points defined at the iteration level will be used insteadof (override) any data configuration on the deployment schedule page. Every new iteration added willinherit this value. You can customize this value for each iteration, if needed.

Note: If fewer data points exist within the time range configured on the deployment schedule pagethan the number of data points specified in the step or iteration level, the available data points will beused. The system will not interpolate to fill the missing data points.

The list of inputs and outputs appear on the new iteration where you can map tags in the Asset Taglist or Asset Hierarchy list.

10. Map the inputs and outputs normally.11. Select Save Changes.

Manage Deployment Schedules

About Analytics Deployment SchedulesUse the Schedule step to create a schedule for running a deployed analytic or execute it on-demand.

You can run the analytic on demand, set the analytic to run continuously, define an end date and time, orstop the run manually. You can select the following units of time and frequency for a recurrent schedule:

Table 7: Execution Time Options

Unit Range Applicable to

Seconds Select a numerical value from 1 to 59. • Repeats Every• Sample Duration (Batch Size)• Sampling Interval

Minutes Select a numerical value from 1 to 59. • Repeats Every• Sample Duration (Batch Size)• Sampling Interval

Hours Select a numerical value from 1 to 23. • Repeats Every• Offset• Sample Duration (Batch Size)

68 © 2020 General Electric Company

Page 73: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Unit Range Applicable to

Days Select one day. • Offset• Sample Duration (Batch Size)

Samples Select a numerical value for the numberof samples before the offset.

• Sample Duration (Batch Size)

For each data request, you must define how much data will be pulled and the interval between datapoints.

Table 8: Data Requests

This request Means this

Offset The amount of time in the past to begin the data pull.

Sample Duration (Batch Size) The amount of data being pulled before the Offset.

Time Span The time span for which the data is being pulled.

Sampling Interval The interval between data points.

Schedule a Recurrent Analytic

You can set an analytic to run continuously, define how often it runs, define how far back in time the datais requested, define how large the sample is and how often the data is sampled.

Before You Begin

Make sure that you have selected applicable assets, and mapped input and output tags for deployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. In the Search pane, select an existing deployment to access its workspace.

3. In the page heading, select Edit to access the 1. Asset Selection section.

4. Select the 3. Schedule tab.

5. In the SCHEDULE subsection, select Recurrent.

6. Enter a value in the Repeats Every box, and select a time unit from the drop-down list box.

7. In the DATA REQUEST subsection, enter a value in the Offset Before Schedule box, and then select atime unit from the drop-down list box. The permitted values for offset time is hours, days, months, andyears.

The offset determines how long ago the data request should begin.

Note: Entering 0 in the Offset field equates to the present day.

8. Enter a value in the Priority box.

The sample duration determines how much data is pulled.

9. Enter a value in the Sampling Interval field, and then select a time unit from the drop-down list.

The sampling interval determines the interval between data points.

10. Select Save to save the scheduling configuration.

11. Select Deploy.

Deploy is not enabled unless all conditions are met.

Tip: Do not select Deploy until you have reviewed the deployment configuration. If you modify theschedule, you must redeploy the analytic to activate these changes.

© 2020 General Electric Company 69

Page 74: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

The analytic is deployed to the respective runtime environment. After deployment, the analytic runswhen scheduled.

Next Steps

Review the deployment.

Schedule an Analytic On Demand

You can schedule an analytic to run on-demand and configure how the data is sampled.

Before You Begin

Make sure that you have selected applicable assets, and mapped input and output tags for deployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.

2. In the Search pane, select an existing deployment to access its workspace.

3. In the page header, select Edit to access the 1. Asset Selection section.

4. Select the 3. Schedule tab.

5. In the SCHEDULE subsection, select Only Once.The Time Span date picker appears under DATA REQUEST subsection. The default time span is oneday.

6. Make the start and end time span selections in the Time Span date picker.

7. Enter a value in the Priority box.

8. Select Save to save the scheduling configuration.

9. Select Deploy.

Deploy is not enabled unless all conditions are met.

Tip: Do not select Deploy until you have reviewed the deployment configuration. If you modify theschedule, you must redeploy the analytic to activate these changes.

The analytic runs in the background.

Next Steps

Review the deployment.

Schedule a Streaming Analytic

Tenant-specific option to deploy and run analytics for continuous streaming.

Before You Begin

This procedure assumes that the following prerequisite tasks have been completed.

• You have uploaded the required assets, tags and time series data required for this analytic.• Your tenant has been configured to stream to Predix timeseries.• You have the required access permissions to manage analytics on Spark runtime for the OPM tenant.• You have added and configured a streaming analytic in the analytic catalog.• You have created a deployment for that streaming analytics.• You have applied the deployment to the applicable assets through asset selection.• You have mapped tags to the inputs, constants and outputs before deploying the analytic job to the

Spark runtime environment

70 © 2020 General Electric Company

Page 75: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

About This Task

The streaming schedule is only available for tenants that are configured to support big data streamingthrough Spark runtime cluster environment. Depending on the tenant configuration, this option may notbe available. Use the Schedule section to deploy and run your forecasting and streaming analytic jobs.You can run the deployed job on demand or through continuous streaming. Once started, the streamingjob runs continuously.

Once you have selected the tags to map, you can continue to the 4. Review tab, or save and close.Navigating by selecting outside the deployment screen saves and closes the access deployment. Selectthe Deployment Name link to continue your deployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page header, select Edit to access the 1. Asset Selection section.4. Select the 3. Schedule tab.5. In the SCHEDULE subsection, select Streaming.6. Enter a value in the Sampling Interval box, and then select Seconds or Minutes in the drop-down list

box. For example, to fetch a data point for every minute within the specified date/time range, enter 1,and then select the denomination as minute.

7. Select Save to save the scheduling configuration.8. Select Deploy.

Deploy is not enabled unless all conditions are met.9. Select Close to save and close the deployment without deploying.

Results

After deployment, the analytic is initiated in the Spark runtime environment and runs continuously until itis manually stopped.

Schedule an Orchestration

You can schedule an orchestration to run at regular intervals, or run the orchestration on demand.

About This Task

Similar to any analytic deployment, you can also use the 3. Schedule section of the deployment to runorchestrations.

Note: If the user has configured the Data Points value/field for a particular analytic step or and/or any ofits iterations, these data point values will be used (override) instead of any data-related settings inscheduling a recurrent orchestration.

The following are the Orchestration Scheduling Enhancements:

• Data request parameters can be set on per step basis. Data request parameters, such as Start Time,End Time, for Run Once, Offset before schedule, and Sample duration, which are set at orchestrationlevel, apply to all steps by default. However, you can set the parameters on per step basis and overridethese parameters.

• In addition to the Historian mode, the following modes are available in the Time Series data:

◦ Math◦ Sample and Hold

• Runtime supported for Predix Insights only.

© 2020 General Electric Company 71

Page 76: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

You can run the orchestration on demand, set the orchestration to run continuously, and define an enddate and time. You can select the following units of time and frequency for a recurrent schedule:

Table 9: Execution Time Options

Unit Range Applies to the following fields

Seconds 1-59 • Repeats Every

• Sample Duration (Batch Size)

• Sampling Interval

Minutes 1-59 • Repeats Every

• Sample Duration (Batch Size)

• Sampling Interval

Hours 1-24 • Repeats Every

• Offset

• Sample Duration (Batch Size)

Days Select one or more days of the week. • Offset

• Sample Duration (Batch Size)

Months 1-12 Offset

Years Select a numerical value. Offset

Samples Select a numerical value for the number

of samples before the offset.

Sample Duration (Batch Size)

For each data request, you must define how much data will be pulled and the interval between datapoints.

Table 10: Data Requests

This request Means this

Offset The amount of time in the past to begin the data pull.

Sample Duration (Batch Size) The amount of data being sampled before the Offset.

Time Span The time span for which the data is being sampled.

Sampling Interval The interval between data points.

Procedure

• Schedule a recurrent analytic.• Schedule an analytic on-demand.

Schedule a Analytic to Trigger on an Event

Subject to your tenant configuration, you can trigger an analytic to run when a specific alert eventhappens.

Before You Begin

This procedure assumes that the following prerequisite tasks have been completed.

72 © 2020 General Electric Company

Page 77: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

• You have uploaded the required assets, tags, and time series data required for this analytic.• You have alert templates defined for your tenant.• Your tenant has been configured to enable the PowerDataFabric runtime.• You have the required access permissions to manage analytics.• You have added and configured an analytic template for the PowerDataFabric runtime in the analytic

catalog.• You have created a deployment for that analytic template.• You have applied the deployment to the applicable assets through asset selection.• You have mapped tags to the inputs, constants, and outputs before deploying the analytic job.

About This Task

The triggering schedule is only available for tenants that are configured to support event trigger throughthe PowerDataFabric runtime environment. Depending on the tenant configuration, this option may notbe available. Use the Schedule section to deploy and trigger analytic jobs. You can run the deployed jobon demand or through event trigger. Once started, the job is only triggered when one or more of theconfigured alert event happens.

Once you have selected the tags to map, you can continue to the 4. Review section, or save and close.Navigating by selecting outside the deployment screen saves and closes the deployment.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page heading, select Edit to access the 1. Asset Selection section.4. Select the 3. Schedule tab.5. In the SCHEDULE subsection, select Triggered.6. In the Alert Templates subsection, add alert templates.

Important: You must not add alert templates that are configured in the analytic template used in thecurrent deployment. This will cause a circular dependency, and the job will fail.

a) In the TEMPLATE column, add the alert template.b) In the EXECUTION DELAY column, add the time delay needed to trigger the job to run.c) In the START OFFSET column, add the time offset before the job starts.d) In the END OFFSET column, add the time offset after the job ends.

7. In the DATA REQUEST subsection, add the following:Option

Offset Before Schedule Enter a value in the box, and then select a time unit in thedrop-down list box. The permitted values for offset time ishours, days, months, and years.

Sample Duration (Batch Size) Enter the amount of data to be sampled before the Offset.

The offset determines how long ago the data request should begin.

Note: Entering 0 in the Offset box equates to the present day.8. Enter a value in the Sampling Interval box, and then select Seconds or Minutes in the drop-down

list box. For example, to fetch a data point for every minute within the specified date/time range, enter1, and then select the denomination as minute.

9. Select Save to save the scheduling configuration.10. Select Deploy.

The Deploy button is not enabled unless all conditions are met.

© 2020 General Electric Company 73

Page 78: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

11. Select Close to save and close the deployment without deploying.

Results

After deployment, the analytic is initiated in the PowerDataFabric runtime environment whenever theconfigured alert events happen.

Review Deployments

Review Deployment ConfigurationUse the Review step to access, modify, and deploy your selections.

Before You Begin

Make sure you have completed the first three steps of deployment.

About This Task

Each subsection in the Review section includes information about your deployment configuration efforts.After completing all the required configuration modifications, you can deploy the analytic to therespective runtime environment, and run it at the scheduled time.

Procedure

1. In the module navigation menu, go to Analytics > Deployments.2. In the Search pane, select an existing deployment to access its workspace.3. In the page header, select Edit to access the 1. Asset Selection section.4. Select the 4. Review tab.5. Make the required modifications to relevant subsections of the page. Do one or more of the following:

• Make required modifications, then select Save in the step before proceeding to the next step.• Select Next to proceed to the next section without modifying the current section.

After modifying the required subsections, return to the 4. Review section when you are done.6. Select Deploy to upload the deployment to the runtime and run the analytic.

When you select Deploy or Close, you are returned to the Deployments details page.

74 © 2020 General Electric Company

Page 79: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

8Analytics TroubleshootingTopics:

• Cannot Access ApplicationAnalytics Module and Apps

• Cannot Access ApplicationAnalytics Module and/orFeatures

• Cannot Access CustomProperties For Alerts in AnalyticOutput

• Cannot Deploy an Analytic toPredix Runtime

• Cannot Access Attributes in theAsset Filter

• Failed Analytics Deployment

© 2020 General Electric Company 75

Page 80: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Cannot Access Application Analytics Module and AppsYou cannot access the Analytics module and apps in your tenant.

Condition

You are logged in to your tenant with the necessary privileges to access the Analytics module and apps.However, when you select any navigation links in the Analytics module, the following error messageappears: Your tenant provisioning is incomplete. Request your tenantAdministrator to create the analytic functional user in the Tenant.

Cause

Your tenant administrator has not set up the password for the analytic system user to enable theApplication Analytics to communicate with the configured runtime for your tenant.

Remedy

Perform the steps in Set Up the Analytics Functional User on page 4.

Cannot Access Application Analytics Module and/or FeaturesYou are not able to access some or all of Application Analytics modules, sub-modules, apps and featuresin your tenant.

Condition

You are logged in to your tenant, you have one or more of the following problems:

• You cannot access any sub-modules.• You can access but cannot make additions or modifications in the sub-module/app. For example, you

can access deployments, but cannot add new deployments or modify existing deployments.• You cannot access some of the sub-modules/apps. For example, you can view Analytics Templates and

Orchestrations, but cannot access Deployments or Asset Filters.

Cause

Your tenant administrator has not enabled access to the required permission sets to access theappropriate sub-modules/apps. Your tenant configuration may be older and still uses the wide defaultpermission set of Manage Analytics Catalog. It must be configured with the granular permission sets.

Remedy

Work with your tenant administrator to give you the needed permission sets to access the required sub-modules/Apps.

76 © 2020 General Electric Company

Page 81: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Cannot Access Custom Properties For Alerts in Analytic OutputYou cannot access the custom properties key-values pairs that were passed to GE Digital APM Alertsthrough an analytic output event.

Condition

The GE Digital APM alert service does not publish the custom properties passed through events in theanalytic output. The analytic had passed these custom properties as an array of key-value pairs in theanalytic output.

Cause

The related keys in custom properties of an analytic output are not pre-registered within the specifictenant.

Remedy

Because of security considerations, prior to passing values as alert attributes via the custom properties ofan analytic output, you are required to first register all related keys in custom properties within thespecific tenant. This step allows the tenant to recognize these keys as valid, and properly pass the key-values to an alert as an attribute, allowing access and visualization in Alerts via the grid view. To registerkey(s) as custom attribute(s) recognizable by GE Digital APM, follow these steps:

1. Determine all keys to be registered.2. Register all keys per instructions in the API documentation. Select the link below to access the API

documentation:https://apm-apidocs-preprod.preprod-app-api.aws-usw02-pr.predix.io/alarms/#_create_custom_attribute

3. After you have registered all keys, pass the key-value pairs for alert custom properties in the requiredJSON format of the analytic output.

Related Information

#unique_107

Cannot Deploy an Analytic to Predix RuntimeYou cannot deploy an analytic even after completing all the required steps in deployment.

Condition

You have uploaded an analytic template and configured all required inputs, outputs, and constants. Yousuccessfully add a deployment with the configured Analytic Template. However, after completing all thesteps required in the deployment successfully, the error appears when you select Deploy. Thedeployment status remains Not Deployed.

Cause

The analytic template has not yet been uploaded to the analytic catalog. The current status of the analytictemplate still displays In Progress. Deployments cannot be staged to the Predix runtime unless theanalytics template is available in the catalog.

© 2020 General Electric Company 77

Page 82: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Remedy

Make sure that the current status of the analytic template used in your deployment no longer displays asIn Progress. Open the deployment in the edit mode. Navigate to the Review page, and then selectDeploy to stage the deployment to Predix runtime.

Verification

The deployment status should changes from Not Deployed to In Progress. This indicates that thedeployment has been uploaded to the runtime to be staged for execution.

Cannot Access Attributes in the Asset FilterIn the Analytics module, you cannot access a list of attributes during asset filter creation.

Condition

You are logged in to your tenant with the necessary privileges to access the Analytics module and apps.However, when you are adding a new asset filter, you are not able to see all required attributes to createthe filter.

Cause

You have not yet configured the attribute, or you have configured attributes that are not part of your assetmodel. Using non-existing model attributes in your analytics configuration will not add them to the assetmodel, nor will they show up for use in the asset filters as they do not exist in the system.

Remedy

Add the needed system attributes to the asset model in your tenant. Use system attributes that you haveingested in your analytics configuration. After, you have configured them, you can use them in the assetfilters.

Failed Analytics DeploymentYou cannot run the deployment successfully on Predix runtime.

Condition

You selected the deploy button to deploy an analytic to run on the Predix runtime. You had completed theasset selection, I/O mapping, scheduling, and review steps prior to deploying the analytic to the runtime.However, after running the deployment you see the following error message when you hover over theRun Once Failed deployment status:

“Invalid input port name <analyticInputName>. Cannot find it in ports to data source list.”

Cause

It is possible that the analytic generates alerts but the output event is not defined in your deployedanalytic template. This causes the alert ingestion to fail as no template definition can be found for it.

78 © 2020 General Electric Company

Page 83: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Remedy

Make sure that you have created alert templates and linked these templates as output events in youranalytic template configuration. This will ensure that any alert generated by the deployed analytic will bewritten to the correct alert template.

© 2020 General Electric Company 79

Page 84: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

9ReferenceTopics:

• Analytics Developer Reference• Sample Analytics• Understand Analytic Ingestion

80 © 2020 General Electric Company

Page 85: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Analytics Developer Reference

About Analytics Developer DocumentationGE Digital APM includes some developer documentation to bridge the gap between Predix platformdocumentation and GE Digital APM user documentation.

Developer documentation included here:

• Clarifies preparation work for privileged users.• Adds information and procedures for back-end functionality.• Provides reference material for developers when configuring tenant provisioning.• Provides reference material for developers when using GE Digital APM APIs.

Sample Data Map for Analytic Template Definition

The data map shows the correlation between the analytic template definition in the user interface andthe analytic inputs and outputs.

At least one input is required. An analytic can optionally define outputs, events, or both. Inputs, outputs,and constants in the analytic template definition must match the input parameters, output parameters,and constants specified in the developed analytic to enable its successful deployment.

Template input definitions and constants correlate with the JSON body of inputs:

© 2020 General Electric Company 81

Page 86: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

The analytic must follow the JSON contract prescribed by the respective runtime environment andlanguage for succesfully running the analytic and generating the output. Template output definitions andconstants correlate with the JSON body of outputs:

Output events in the analytic template definition map to an existing alert template or the alert templateprovided as part of the template definition:

Config.json information for developing analytics can be found in the sub-topics in the related link below.

Analytic Definition Files

You can create .csv files containing the relevant inputs, constants, and outputs for your analytic instead ofentering them manually in the analytic template. To do this, you must save a copy of the .csv template andadd the appropriate definitions before uploading it to the analytic template. If you upload a file with thewrong format or incorrect values, a dialog box displaying the invalid entries appears, and the file is notloaded. You can fix the indicated errors, and retry.

82 © 2020 General Electric Company

Page 87: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Tip: You can use the apm-analytic-csv-generator found on GitHub if you have access. You canrequest access from your support team. The .csv generator parses a sample .json analytic input file andwrites out input and constant .csv files.

Related Information

https://github.build.ge.com/212348840/apm-analytic-csv-generator

Define Analytic Input CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple inputs through .csv upload. You can skip thisprocedure if you want to add inputs directly through the UI.

Configure the template to specify input tags for the analytic you are adding or have added to the analyticcatalog. This information is needed to run the analytic.

Procedure

1. In the Analytic Template section, select in the Input Definitionsubsection to download a copy ofthe .csv template and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the AnalyticInputDefTemplate.csv file as AnalyticInputDef_data_wind_trend.csv and save it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values.

Table 11: Input Definition Guidelines

This definition Uses this guideline

Name A valid tag name that matches the analytic input parameter. Itmust match the regular expression pattern ^[A-Za-z0-9-_]+$.

Description The tag description. If there is no description for the tag,define blank entries as empty table cells, or offset them in acomma-delimited set (, ,).

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING

Note: You must select string data type for data of typestring, character, enum, or grid.

• INTEGER• DOUBLE• BOOLEAN

© 2020 General Electric Company 83

Page 88: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

This definition Uses this guideline

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the tags and attributes of allassets and their children starting at the defined Entity Type.For example, if you configure the Entity Type for your analyticinput at the Site level, then all tags and attributes of the Siteand its children become available for mapping duringdeployment.

Required The value must be checked or unchecked. Thisindicates whether the input is required.

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.Define Analytic Constants CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple constants through .csv upload. You can skip thisprocedure if you want to add constants on the UI.

Configure the template to specify constants for the analytics you are adding or have added to the analyticcatalog. This information is needed to run the analytics.

Procedure

1. In the Analytic Template section, select in the Constantssubsection to download a copy ofthe .csv template and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the AnalyticConstantsTemplate.csv file asAnalyticConstantsTemplate_data_wind_trend.csv and save it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values.

Table 12: Constants Guidelines

This definition Uses this guideline

Name A valid tag name must match the analytic constantparameter. It must match the regular expression pattern^[A-Za-z0-9-_]+$.

Description The tag description or null. If there is no description for thetag, define blank entries as empty table cells, or offset them ina comma-delimited set (, ,).

84 © 2020 General Electric Company

Page 89: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

This definition Uses this guideline

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING• INTEGER• DOUBLE• BOOLEAN

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the attributes of all assetsand their children starting at the defined Entity Type. Forexample, if you configure the Entity Type for your analyticconstant at the Site level, then all attributes of the Site and itschildren become available for mapping during deployment.

Value (optional) If there is no default value, you must define blank entries asempty table cells or offset them in a comma-delimited set(, , ).

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.Define Analytic Output CSV

Before You Begin

You have created an analytic in the analytic catalog.

About This Task

Use this procedure if you want to configure multiple outputs through .csv upload. You can skip thisprocedure if you want to add outputs on the UI.

Configure the template to specify output tags for the analytic you are adding or have added to theanalytic catalog. This information is needed to run the analytic.

Procedure

1. In the Analytic Template section, select in the Outputssubsection to download a copy of the .csvtemplate and save it locally.

2. Rename the file with a name related to the analytic.For example, rename the output-def.csv file as output-def_data_wind_trend.csv andsave it locally.

3. Open the saved file in a CSV editor.4. Verify that the template matches the following definition guidelines. They can be defined in

spreadsheet columns or as comma-separated values. Make sure your outputs match your analyticdefinition.

© 2020 General Electric Company 85

Page 90: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 13: Outputs Definition Guidelines

This definition Uses this guideline

Name A valid tag name that matches the analytic constantparameter. It must match the regular expression pattern^[A-Za-z0-9-_]+$.

Description The tag description or null. If there is no description for thetag, define blank entries as empty table cells, or offset them ina comma-delimited set (, ,).

Units The engineering unit that matches the engineering unit namein the conversion database.

Data Type The value must be one of the following data types:

• STRING• INTEGER• DOUBLE• BOOLEAN

Data Format The format of the data being retrieved or written as part ofthe analytics computation logic. The two acceptable valuesare Asset Attributes and APMTimeseries.

Entity Type The mapping level in the asset business hierarchy. During theI/O mapping step in deployment, the mapping level tellsApplication Analytics to retrieve the tags and attributes of allassets and their children starting at the defined Entity Type.For example, if you configure the Entity Type for your analyticoutput at the Segment level, then all tags and attributes ofthe Segments and its children become available for mappingduring deployment.

5. Make any required modifications and save the file.

Next Steps

Upload the template to the new analytic created in the catalog.

Generate an Alert from an Orchestration

You can generate alerts based on the analytic output from an orchestration associated with a specificasset.

Configure Tags for Alert Generation

The following steps represent a typical end-to-end workflow of how the system generates alerts from ananalytic orchestration output:

1. Configure or use an existing Analysis View.2. Configure or use an existing Alert Template.3. Link an Asset Analysis View to an Alert Template.4. Configure and deploy an analytic. Use the Alert Template as the output event.5. Configure an Asset Filter for selection in the orchestration.6. Add an orchestration. Use the Asset Filter for applicability in the orchestration.7. Use the deployed analytic in the Analytic Data Flow in the orchestration.8. During I/O Mapping, select and map the new calculated tag.9. Create a deployment for the orchestration.10. You can run the orchestration using one of three methods:

86 © 2020 General Electric Company

Page 91: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

• On-Demand Run.• Scheduled Recurrent Runs.• Event-Based Triggering.

Refer to Scheduling an Orchestration and Event-Based Triggering of an Analytic Execution Using anExternal Packet (Data Set).

Analytic Orchestartion Generated Events and Alerts

The output of the analytic orchestration is stored in time series with a unique tag name and identifier. Thealert services supports publishing of common properties and custom properties for output events in ananalytic output. These properties are an array of key-value pairs that can be used to supplement the alertinformation. GE Digital APM uses these output values to generate alerts, some of which can then beaccessed via Alerts and other modules.

The following applies to analytic generated alerts:

• GE Digital APM uses the commonProperties in the analytic's ouput alert block for each alert event.• The commonProperties values defined within the event itself, takes precedence over the values

defined in the commonProperties block at the root level. The commonProperties key-valuesgenerated as part of an analytic input cannot be currently visualized in the Alerts UI.

• The template key-value defined within the output event, takes precedence over the ouput eventdefinition in the analytic template.

{"date":40000,"score":5.2,"sensors":["S1","S2"],"template":"customAlertTemplate"}

• The templateSuffix key-value gets appended to the base template.

{"date":50000,"score":10.2,"sensors":["S1","S2"],"templateSuffix":"_High"}

• A null override removes the commonProperties from a specific event. A blank override for a keywould save an empty string value for that key.

Table 14: Supported Data Types for Alert Generation

Data Source Data Type

Inputs (for the analytic) Time series

Constants • For Expression: String

• For Value: Double

Orchestration Output (single-step orchestration) • JSON Format

• Time series values as output

GE Digital APM Alerts • GE Digital APM Alert objects

• Based upon alert template configuration in the output

event of the analytic template, or the custom alert

template specified in the analytic output payload for the

event.

When an analytic is deployed, a tagType and tag association is created for each asset if such tagType andtag association does not exist. The tag type name is created using the rule <analytic-name>_<output-name>. The tag association name is created using the rule <asset-source-key>_<analytic-name>_<deploy-cfg-name>_<output-name>.

© 2020 General Electric Company 87

Page 92: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Pre-Registering Custom Property Keys

For security considerations, prior to passing values as alert attributes via the custom properties of ananalytic output, your are required to first register all related keys in custom properties within the specifictenant. This step allows the tenant to recognize these keys as valid, and properly pass the key-values toan alert as an attribute, allowing access and visualization in GE Digital APM Alerts via the grid view. Toregister key(s) as custom attribute(s) recognizable by GE Digital APM, follow these steps.

1. Determine all keys to be registered.2. Register all keys per instructions in the API documentation. Select the link below to access the API

documentation:https://apm-apidocs-preprod.preprod-app-api.aws-usw02-pr.predix.io/alarms/#_create_custom_attribute

3. After you have registered all keys, follow the documentation below, regarding the JSON format andoutput.

Sample Input JSON (Request to Orchestration Engine)

[ { "assetUri": "/assets/ddb90c85-64af-386d-a792-1b8733a63369", "assetSourceKey": "56000", "assetName": "56000", "messageDateTime": 1472755021693, "assetFilterFields": [ { "name": "ENGINE_FAMILY", "value": [ "CF34" ], "dataType": "String", "priority": 1 }, { "name": "ENGINE_TYPE", "value": [ "CF34" ], "dataType": "String", "priority": 2 }, { "name": "AIRCRAFT_FAMILY", "value": [ "B737" ], "dataType": "String", "priority": 3 }, { "name": "AIRCRAFT_TYPE", "value": [ "B737-800" ], "dataType": "String", "priority": 4 },

88 © 2020 General Electric Company

Page 93: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

{ "name": "CARRIER_CODE", "value": [ "DELTA" ], "dataType": "String", "priority": 5 } ] } ]

Alert Outputs

The following example payload represents the alert block of an analytic deployed in the Predix runtime.

{"alerts":{"commonProperties":{"prop1":"val1","prop2":"val2"},"events":[{"date":10000,"score":2.4,"sensors":["S1","S2"]},{"date":20000,"score":3.1,"sensors":["S1","S2"]},{"date":30000,"score":4.2,"sensors":["S1","S2"]},{"date":40000,"score":5.2,"sensors":["S1","S2"],"template":"customAlertTemplate"},{"date":50000,"score":10.2,"sensors":["S1","S2"],"templateSuffix":"_High"},{"date":70000,"score":7.3,"sensors":["S1"],"prop1":"override"}]},"time_series":{"mean":[20, 30],"time_stamp":[10000,20000,30000]}}

The following example payload is what the system sends to the Alert.

[{"name":<Template Name from IO Map>,"eventStart":10000,"scanGroupData":{"scores":2.4},"storageReceiveTime":<Now>,"severity":<Severity from Template>"tagsOfInterest":["S1","S2"],"prop1":"val1","prop2":"val2"},{"name":<Template Name from IO Map>,"eventStart":20000,"scanGroupData":{"scores":3.1},"storageReceiveTime":<Now>,"severity":<Severity from Template>"tagsOfInterest":["S1","S2"],"prop1":"val1","prop2":"val2"},

© 2020 General Electric Company 89

Page 94: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

{"name":<Template Name from IO Map>,"eventStart":30000,"scanGroupData":{"scores":4.2},"storageReceiveTime":<Now>,"severity":<Severity from Template>"tagsOfInterest":["S1","S2"],"prop1":"val1","prop2":"val2"},{"name":"customAlertTemplate","eventStart":40000,"scanGroupData":{"scores":5.2},"storageReceiveTime":<Now>,"severity":<Severity from customAlertTemplate>"tagsOfInterest":["S1","S2"],"prop1":"val1","prop2":"val2"},{"name":<Template Name from IO Map>_High,"eventStart":50000,"scanGroupData":{"scores":10.2},"storageReceiveTime":<Now>,"severity":<Severity from <Template IO Map>_High>"tagsOfInterest":["S1","S2"],"prop1":"val1","prop2":"val2"},{"name":<Template Name from IO Map>,"eventStart":70000,"scanGroupData":{"scores":7.3},"storageReceiveTime":<Now>,"severity":<Severity from Template>"tagsOfInterest":["S1"],"prop1":"override","prop2":"val2"}]

Output JSON

{ "uri": "/schedules/212a49f6-af98-45cf-9403-37511c21f33f", "createdOn": 1480458946415, "updatedOn": 1480465404146, "orchestrationUri": "/orchestrations/578d2b94-3a87-4f2c-904c-5b6d87bb8c08", "assetUri": "/assets/99009f66-67e0-3708-b8d0-75872ec19060", "job": { "id": "8cc7f428-14a4-4eee-9427-c82e8239da80", "name": "Auto1480455130446d1eae08f-7317-4ee0-b0d2-e09c7f8316ea", "description": "/orchestrations/578d2b94-3a87-4f2c-904c-5b6d87bb8c08, 1303e8a5-f81a-4e65-b7ac-abd091459fcc, /assets/99009f66-67e0-3708-b8d0-75872ec19060", "cron": {

90 © 2020 General Electric Company

Page 95: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"seconds": "0", "minutes": "0/7", "hours": "*", "dayOfMonth": "?", "months": "*", "dayOfWeek": "*", "years": "*", "timeZoneId": "UTC" }, "state": "Active", "executionRequest": { "url": "https://predix-analytics-execution-release.run.asv-pr.ice.predix.io/api/v2/execution/async", "httpMethod": "POST", "httpHeaders": [ { "name": "Content-Type", "value": "application/json" }, { "name": "Predix-Zone-Id", "value": "5f3fd009-dd9a-41d1-b7a2-4c34a0931237" } ], "inputData": "{\"orchestrationConfigurationId\":\"1303e8a5-f81a-4e65-b7ac-abd091459fcc\",\"assetId\":\"/assets/99009f66-67e0-3708-b8d0-75872ec19060\",\"dataSource\":[]}" }, "createdBy": "cg5t5bm878qhwofouta7r6jgw4o29hmmnp8m", "updatedBy": "cg5t5bm878qhwofouta7r6jgw4o29hmmnp8m", "createdTimestamp": "2016-11-29 22:35:46.39", "updatedTimestamp": "2016-11-29 23:25:34.555" }, "historyCheckedTime": 1480465398362, "jobHistory": [ { "jobEvent": { "id": "bb3d2947-bdf4-4c69-a502-da8dae5295c4", "jobId": "8cc7f428-14a4-4eee-9427-c82e8239da80", "fireTime": "2016-11-29 22:42:01.49", "scheduledFireTime": "2016-11-29 22:42:00.0", "cron": "0 0/3 * ? * * * ", "timeZoneId": "UTC", "httpMethod": "POST", "url": "https://predix-analytics-execution-release.run.asv-pr.ice.predix.io/api/v2/execution/async", "httpStatusCode": 200, "statusMessage": "completed", "result": "{\"orchestrationExecutionStatus\":[{\"assetId\":\"/assets/99009f66-67e0-3708-b8d0-75872ec19060\",\"errorResponse\":null,\"orchestrationStepStatus\":[],\"contextId\":\"09407eb3-5c9e-4592-93c2-5075e2f26fe0-/assets/99009f66-67e0-3708-b8d0-75872ec19060\",\"startTime\":null,\"endTime\":null,\"status\":\"PROCESSING\"}],\"errorResponse\":null,\"orchestrationRequestId\":\"09407eb3-5c9e-4592-93c2-5075e2f26fe0\",\"status\":\"PROCESSING\"}" }, "executionOutput": { "executionResult": { "orchestrationRequestId":

© 2020 General Electric Company 91

Page 96: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"09407eb3-5c9e-4592-93c2-5075e2f26fe0", "status": "COMPLETED", "orchestrationExecutionStatus": [ { "contextId": "09407eb3-5c9e-4592-93c2-5075e2f26fe0-/assets/99009f66-67e0-3708-b8d0-75872ec19060", "assetId": "/assets/99009f66-67e0-3708-b8d0-75872ec19060", "status": "COMPLETED", "orchestrationStepStatus": [ { "status": "COMPLETED", "analyticId": "2038d257-2e98-486c-832d-ca1ae39b6704", "analyticName": "Auto1480455130446", "analyticVersion": "1.8", "analyticRequestId": "0ddbc189-b685-11e6-8552-2ae52e760e7c-sid-c6b1bbb5-deb9-4349-8e13-9a3e82c5bd83", "stepId": "sid-c6b1bbb5-deb9-4349-8e13-9a3e82c5bd83", "startTime": 1480459326140, "endTime": 1480459340250, "output": "{"alerts":{ "commonProperties":{"prop1":"val1","prop2":"val2"}, "events":[ {"date":10000,"score":2.4,"sensors":["S1","S2"]}, {"date":20000,"score":3.1,"sensors":["S1","S2"]}, {"date":30000,"score":4.2,"sensors":["S1","S2"]}, {"date":40000,"score":5.2,"sensors":["S1","S2"],"template":"customAlertTemplate"}, {"date":50000,"score":10.2,"sensors":["S1","S2"],"templateSuffix":"_High"}, {"date":70000,"score":7.3,"sensors":["S1"],"prop1":"override"} ] }, "time_series":{ "mean":[20, 30], "time_stamp":[10000,20000,30000] } }", "errorResponse": null } ], "startTime": 1480459326080, "endTime": 1480459340250, "errorResponse": null } ], "errorResponse": null }, "asyncExecuteRequestUri": null, "alertsCreated": true, "resultDeletedFromCache": true, "isAlertsCreated": true }, "alertCreateError": null }

92 © 2020 General Electric Company

Page 97: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Accessing Generated Tags

The newly created tag and its output value can be used to automatically generate alerts, which can thenbe accessed in an Alert Inbox. Alerts are generated based on the criteria configured for inputs, outputs,and constants in the analytic template. To access the generated Alert in the user interface, the user mustlink an Alert Template to the Analytic Template. The user will then be able to access the Alert in the userinterface for dispositioning and review in other areas. Alerts may require several minutes to generatefollowing completion of an analytic orchestration execution. Alerts automatically create an Alert Analysisthat can be accessed and modified in Analysis. Refer to About Alerts and its child topics to address anyalerts.

To Access an Alert Analysis, refer to Accessing an Alert Analysis Chart.

To access the tag in a dashboard, you will need to configure a dashboard, card and widget for the selectedasset. Refer to Adding, Modifying, and Deleting Dashboards, Adding and Deleting Cards on a Dashboard,and Adding and Deleting Widgets on a Card. When you configure a selected widget, you can use the newlycreated tag.

Refer to Configure Widgets and its child topics.

Tip: You can set up a widget to navigate to another dashboard such as Alerts or Cases, or to Analysischart.

Refer to Navigating from a Widget or a specific widget under Configure Widgets.

Event-Based Triggering of an Analytic Execution Using an External Packet (Data Set)

You can implement a REST API to trigger orchestration runs from external data packets.

Use the following REST API to retrieve external data packets, trigger orchestrations, and then process theanalytic output data.

One or more data sets that are stored in an external system are provided to the system for processing.This event allows you to trigger an analytic orchestration execution. The system processes each datapacket in the order in which it is received.

During processing, the resolveAndExecute API gathers and uses data from the external data packet,and Asset and Time-Series data stores. The orchestrationExecution API is triggered and an outputis achieved. The output may be used elsewhere in GE Digital APM.

You will need the following REST APIs:

Table 15: REST APIs

API Description

resolveAndExecute and

orchestrationExecution• Retrieves the initial data packet.

• Initiates data gathering for Asset and Time-Series.

• Resolves and matches the appropriate orchestration to

execute for the asset or asset groups associated with the

data packet. The resolveAndExecute API calls

the orchestrationExecution API to trigger

execution of the orchestrations appropriate for the event

(receipt of data packet).

• Manages all orchestration execution activity, including

output.

Note: For more information regarding use of this API, see the AsciiDoc set (https://apm-apidocs-rc.run.asv-pr.ice.predix.io/).

© 2020 General Electric Company 93

Page 98: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 16: Supported Data Types for GE Digital APM Processing and Orchestration Execution

Data Source Data Type

External Data Packet String

Asset Data String

Inputs for Analytics/Orchestrations Time-Series

Outputs for Analytics/Orchestrations Time-Series

Constants for Analytics/Orchestrations Numeric

Orchestration Output Time-Series

When an analytic is deployed, a tagType and tag association is created for each asset if such tagType andtag association does not exist. The tag type name is created using the rule <analytic-name>_<output-name>. The tag association name is created using the rule <asset-source-key>_<analytic-name>_<deploy-cfg-name>_<output-name>.

Before You Begin

Make sure you have done the following:

• Set up an Analytic Template.• Set up an Alert Template.• Created an analytic that the orchestration can use.• The APIs can access and read the data from the selected data store.• An external source containing the data packets for GE Digital APM to use is available. This source can

consist of a Postgres table, message queue, or another type that can serve to provide data forpackets.

The data packets must have the following JSON structure:

[ { "assetUri": "/assets/ddb90c85-64af-386d-a792-1b8733a63369", "assetSourceKey": "56000", "assetName": "56000", "messageDateTime": 1472755021693, "assetFilterFields": [ { "name": "ENGINE_FAMILY", "value": [ "CF34" ], "dataType": "String", "priority": 1 }, { "name": "ENGINE_TYPE", "value": [ "CF34" ], "dataType": "String", "priority": 2 }, {

94 © 2020 General Electric Company

Page 99: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"name": "AIRCRAFT_FAMILY", "value": [ "B737" ], "dataType": "String", "priority": 3 }, { "name": "AIRCRAFT_TYPE", "value": [ "B737-800" ], "dataType": "String", "priority": 4 }, { "name": "CARRIER_CODE", "value": [ "DELTA" ], "dataType": "String", "priority": 5 } ] } ]

Prepare and Load the Data

You will need to configure at least three data stores to use with the GE Digital APM analytics event-basedtriggering;

• GE Digital APM Asset - to load the asset-related data.• GE Digital APM Time-Series - to load the assets time-series related data.• External data store - to store data packets, or events, which are utilized to initiate analytic

orchestration execution.

1. Load asset related data to the GE Digital APM Asset data store in your tenant.2. Load time-series data (tags) to the GE Digital APM Time-Series data store.3. Configure a data store outside of GE Digital APM to store external data packets.

The individual data packets are considered the events which are utilized to initiate analytic orchestrationexecution. APIs use this external data as the criteria for matching the appropriate Asset and Time-Seriesdata in GE Digital APM for analytic orchestration execution.

Initiate GE Digital APM Processing

To initiate processing, call the resolveAndExecute API for each data packet. This will retrieve datapackets available in the external queue, then access data from the data packet.

After you have uploaded and configured all of the individual Analytics called by the resolveAndExecute API, the system does the following:

1. Extract the appropriate Asset data from the report.2. Link and extract the appropriate Asset, Tag, and Time-Series data from GE Digital APM using existing

data in GE Digital APM Asset and Time-Series.3. Retrieve the appropriate Inputs, Outputs, and Constants for the Orchestration to be executed.

© 2020 General Electric Company 95

Page 100: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

The resolveAndExecute API passes this data to the execution function of GE Digital APM Analytics(orchestrationExecution API) to initiate execution of the analytic and related orchestrations.Analytics are passed to orchestrationExecution one-by-one and are executed individually.

The JSON output of the orchestration is an output “object” containing the status and results of theorchestration.

The specific output of an analytic is a Time-Series data point.

Sample Response from Orchestration Engine

[ { "assetUri": "/assets/6995a24b-937a-30de-be36-e071a426c13d", "assetSourceKey": "56001", "assetName": "56001", "messageDateTime": 1474504115555, "executionOutputs": [ { "executionResult": { "orchestrationRequestId": "3869f0ae-f81d-45b7-a70a-6db8b280c742", "status": "PROCESSING", "orchestrationExecutionStatus":[ { "contextId": "3869f0ae-f81d-45b7-a70a-6db8b280c742-/assets/6995a24b-937a-30de-be36-e071a426c13d", "assetId": "/assets/6995a24b-937a-30de-be36-e071a426c13d", "status": "PROCESSING", "orchestrationStepStatus":[], "startTime": null, "endTime": null, "errorResponse": null } ], "errorResponse": null }, "alertsCreated": false, "resultDeletedFromCache": false, "isAlertsCreated": false }, { "executionResult": { "orchestrationRequestId": "d0475114-8969-4f81-8bbf-ecca7926b89e", "status": "PROCESSING", "orchestrationExecutionStatus": [ { "contextId": "d0475114-8969-4f81-8bbf-ecca7926b89e-/assets/6995a24b-937a-30de-be36-e071a426c13d", "assetId": "/assets/6995a24b-937a-30de-be36-e071a426c13d", "status": "PROCESSING", "orchestrationStepStatus":[],

96 © 2020 General Electric Company

Page 101: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"startTime": null, "endTime": null, "errorResponse": null } ], "errorResponse": null }, "alertsCreated": false, "resultDeletedFromCache": false, "isAlertsCreated": false } ] } ]

Use of Output in GE Digital APM

The specific output of an analytic is a Time-Series data point. Time-Series data points can be used in otherareas of GE Digital APM, such as Analysis.

Use the specified tag in the GE Digital APM application to view the value. For instance, you can create achart in Analysis, or configure a widget in Dashboards.

Alerts can be generated if the results generated meet the criteria configured in the Analytic Template toproduce an alert. An alert with the Alert Template name is available in the Templates module.

Note:

• Refer to Adding and Deleting an Asset Analysis View in the Help documentation to set up an AnalyticTemplate.

• Refer to Creating, Configuring and Deleting an Alert Template and Adding and Removing Analysis Viewin an Alert Template in the Help documentation to set up an Alert Template.

Status and Error Notifications

In passing Analytics to the orchestrationExecution API, you can track the status of orchestrationexecution using two mechanisms:

Sync modeThe output object from the analytic provides a final response to the orchestration execution (Pass,Fail). Use the asciidoc for orchestrationExecution to configure for Sync mode.

Async modeYou will receive a request ID to monitor the status of orchestration execution using the output object.Then you receive the status of each step in the orchestration using the object to monitor the request.An API can be used to pass the request ID to the system for status. Use the asciidoc fororchestrationExecution to configure for Async mode.

© 2020 General Electric Company 97

Page 102: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Sample Analytics

Sample AnalyticsDownload and test sample analytics on the Application Analytics UI.

Note: Use these samples only in a demo environment. These samples are notintended for use in active tenants that are in production use as these contain sampledata only.

Table 17: Different Types of Sample Analytics

Sample Description Sample ZIP Download

Simple Expression Evaluator Analytics Project https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/apm-simple-expression-evaluator-v1.zip

Java Threshold Analytics Project https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/apm-threshold-analytic-java.zip

Matlab Threshold Analytics Project https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/apm-threshold-analytic-matlab.zip

Python Threshold Analytics Project https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/apm-threshold-analytic-python.zip

Sample Output JSON https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/SampleOutput.json

Generate Analytic with Alerts

Steps to upload and deploy a sample analytic with alerts on Predix runtime.

Before You Begin

This procedure assumes that the following prerequisite tasks have been completed.

• You have uploaded the required assets, tags and time series data required for this analytic.• Your tenant's Event Hub is also registered with the GE Digital APM Alert service.• You are signed into an tenant with access privileges to manage analytics on Predix runtime.• You have downloaded the copy of the sample analytics https://apm-application-help-rc.int-app.aws-

usw02-pr.predix.io/apm-threshold-analytic-KVjava-1.0.1-SNAPSHOT.jar• You have downloaded and extracted the copy of the definition files https://apm-application-help-

rc.int-app.aws-usw02-pr.predix.io/AlertKVPair_analytic_definitions.zip locally.

Procedure

1. Add the alert template

a) Select Alert Templates.b) Select + to add new template.c) Enter Analytics in the Alert Template field.d) Select Save.

2. Upload the analytics template to the catalog.

Configure the following information for the analytic.

98 © 2020 General Electric Company

Page 103: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Option

Runtime Predix

Name AlertKVPair_Analytic_Sample

Owner Your Name

Analytic Type Java

Type Version 1.8.

Analytic File https://apm-application-help-rc.int-app.aws-usw02-pr.predix.io/apm-threshold-analytic-KVjava-1.0.1-SNAPSHOT.jar

Analytic Version 1.0.0

Primary Category Monitoring

3. In the Analytic Template, configure the input definition, constant, and output definition through CSVupload and select Save.

4. Enter Analytics in the Output Events field and select Save.5. Add and configure the deployment as follows:

a) Enter AlertKVPair_Deployment in the Deployment Name field and then select Submit.b) In the 1. Asset Selection step, select the asset defined in the analytic, and then select Save.c) Select Next to access the 2. I/O Mapping step.d) Select the Tag drop-down menu and then, select Add Tags....e) In the tag browser, search for the specific temperature tag to map to the input temp_today in the

analytic (for example TAG_Temperature_ID14. After the search displays the tag, drag and dropit onto the input for mapping it.

f) Select Save and Next to save the I/O Mapping configuration.g) In the 3. Schedule step, leave the selection at Only Once for Define how often a new run will be

executed option.h) Select Time Span between starting from when the timeseries data is available for the input tag.

For example, if the starting date is since May of last year to can select May 1, 2017 to currentdate.

i) Leave the Sample Interval at the default value of 1 Minute.j) Select Save and then select Deploy.

The deployment is saved to the Predix runtime. After successful deployment the status updates toRun Once.

6. Claim the analytic alert in the Alerts module.

a) Select Alerts > Unclaimed > .b) Filter alerts by Time Received.

GE Digital APM Port to Field Map for Predix Analytics

An analytics port-to-field map simply creates a mapping derived from the analytic template configurationsuch as input definitions, output definitions, and output events to tell the runtime engine to connect tothe respective data sources to fetch inputs and write outputs. The port-to-field map is itself a JSONstructure.

Type: PortToFieldMap

See the following table for a description of the elements in a PortToFieldMap.

© 2020 General Electric Company 99

Page 104: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Field Description

analyticName Template name as defined in GE Digital APM.

analyticVersion Analytic version provided at the time of analytic upload or

creation.

comment (Optional) Informational only.

orchestrationStepId Applies to analytic orchestration. It is an auto generated ID for

the orchestration step.

iterations Supports multiple iterations for the same analytic. Creates an

entry per iteration.

Example PortToFieldMap JSON

The following example .json represents the port to field map data for per port time series array.

{ "comment": [ "/pxDeployments/23dabd15-ea98-47ba-aec5-c8adb59b630d", "e3692835-411b-4a1a-82cb-c88e73f98c53", "" ], "analyticName": "Shared_timestamp_sample_analytic", "analyticVersion": "1.0.0", "orchestrationStepId": "sid-fb5e38a2-c5fe-4f8a-84d8-2428d7f7361a", "iterations": [ { "id": "0", "inputMaps": [ { "valueSourceType": "DATA_CONNECTOR", "fullyQualifiedPortName": "data.time_series.input1", "fieldId": "Shared_timestamp_sample_analytic_DeploymentStep1_1_input_input1", "queryCriteria": { "start": "${START_TIME}", "end": "${END_TIME}" }, "dataSourceId": "PredixTimeSeries" }, { "valueSourceType": "DATA_CONNECTOR", "fullyQualifiedPortName": "data.time_series.input2", "fieldId": "Shared_timestamp_sample_analytic_DeploymentStep1_1_input_input2", "queryCriteria": { "start": "${START_TIME}", "end": "${END_TIME}" }, "dataSourceId": "PredixTimeSeries" }, { "valueSourceType": "CONSTANT", "fullyQualifiedPortName":

100 © 2020 General Electric Company

Page 105: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"data.constants.threshold", "value": 10 } ], "outputMaps": [ { "fullyQualifiedPortName": "time_series.output1", "fieldId": "Shared_timestamp_sample_analytic_DeploymentStep1_1_output_output1", "dataSourceId": "Temporary,PredixTimeSeries" }, { "fullyQualifiedPortName": "time_series.output2", "fieldId": "Shared_timestamp_sample_analytic_DeploymentStep1_1_output_output2", "dataSourceId": "Temporary,PredixTimeSeries" } ], "inputModelMaps": [] } ]}

Sample Data Map for Analytic Outputs

When you execute a deployment, the system enables you to download a copy of the <source_key>_<asset_name>.json file to your hard drive that contains the analytic json files for every asset that theanalytic runtime ran the analytic:

Table 18: Deployment JSON Files

This JSON file Is related to this

<source_key>_<asset_name>.json Input and Output time series values for a selected asset during

execution.

Note: You can receive multiple files in the ZIP file.

The files analytic.json, deployConfig_Post.json, anddeployConfig_Put_MappingByAsset.json or deployConfig_Put_MappingByTag.jsoncan be used to create an analytic and deploy configuration programmatically, or by using any REST clientsuch as "Postman".

<source_key>_<asset_name>.json

The <source_key>_<asset_name>.json file comprises execution parameters, execution status,time series input provided to the analytic, and alerts and time series data output given by the analytic.There is one <source_key>_<asset_name>.json file for each asset for which the analytic has beenexecuted.

© 2020 General Electric Company 101

Page 106: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 19: Output JSON Structure

JSON

Deployment Metadata "analyticName" : "ThresholdPassthrough4", "analyticUri" : "/analyticEntries/55a6617e-a614-4add-93c4-1b8790341546", "ioMappingUri" : "/ioMappings/014fd31f-a95e-4ea2-a108-d060ad44846d", "entitySourceKey" : "56002", "entityName" : "56002", "ioMappingName" : "Deploy3", "deployStatus" : "Run Once", "filenameToDataObjects" : { }

Ouput Metadata "commonSensors" : [ ], "commonScanGroupData" : { "numtest" : "6.5", "threshold" : "50.0"

Output Alerts "alerts" : { "commonKeyValueStore" : { "templateName" : "CustomKVTestTemplate" }

102 © 2020 General Electric Company

Page 107: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

JSON

Ouput Events "events" : [ { "date" : 1501574800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "105.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501594800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "75.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501600000001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "90.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501614800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "100.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501634800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "175.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501654800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "75.0" }, "templateName" : "CustomKVTestTemplate" }, { "date" : 1501674800001, "keyValueStore" : { }, "sensors" : [ "AV_TAG_1" ], "scanGroupData" : { "scores" : "1160.0" }, "templateName" : "CustomKVTestTemplate" } ]

© 2020 General Electric Company 103

Page 108: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

JSON

Time-series "time_series" : { "time_stamp" : [ 1501574800001, 1501594800001, 1501600000001, 1501614800001, 1501634800001, 1501654800001, 1501674800001 ], "mean" : [ 105.0, 75.0, 90.0, 100.0, 175.0, 75.0, 1160.0 ], "deviation" : [ 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 ] } }

The date, score, and sensor blocks are related. The first elements from these three blocks are used tocreate the first alert; the second elements are used to create the second alert, and so on.

The time_stamp block is related to all other blocks in the time_series block. The first element in the meanblock corresponds to the first element in the time_stamp block; the second element in the mean blockcorresponds to the second element in the time_stamp block, and so on.

For each entry in the mean block, there is one entry in the time series database. The tag id for each datapoint is in the <asset_sourcekey>.<analytic_name>.<output_def_name> format.

For example, for the values [asset_sourcekey = 399, analytic_name = "analytic1",output_def_name = "mean"], the tag id is "3999.analytic1.mean".

The following entries are available in the time series database after the execution foroutput_def_name = "mean", asset_sourcekey = 3999, analytic_name ="analytic1":

104 © 2020 General Electric Company

Page 109: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 20: Time-series Database Entries

tagId timestamp value

3999.analytic1.mean 1430829000000 33.380000000000003

399.analytic1.mean 1430761000000 47.0

... ... ...

Similarly, there are data point entries for the deviation block as well.

Figure 2: Asset JSON Input

The following are possible status and results:

• PROCESSING• DEPLOYED• FAILED

Figure 3: Analytic Deployment Status

Understand Analytic Ingestion

About Analytic Ingestion ServicesApplication Analytics provides specific REST API services to upload or ingest your analytic logic, templatedefinitions, and deployments simultaneously.

You can use either the Analytic UI or any REST client such as Postman to perform the ingestion. However,before you can prepare the .zip file for upload, it is important to understand the configuration, schemadefinitions, and structure for the expected file for upload or ingestion.

GE Digital APM Analytics has three API endpoints, facilitating three different methods of analytic dataingestion.

© 2020 General Electric Company 105

Page 110: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 21: GE Digital APM Analytic Ingestion Service Endpoints

Service Endpoint Description

/analyticEntries/ingestions This facilitates ingesting analytic template configurations only.This requires you to ingest the analytic template, templateconfigurations including supporting files, input, output, andoutput event definitions. This method is currently supportedboth through the UI and service endpoint.

/deployments/ingestions This facilitates ingesting deployment configurations only. Thisassumes that you have already pre-ingested at least oneanalytic template for which you can add and configuredeployments. This requires you to ingest deploymentconfigurations required to run the deployment in the specifiedanalytic runtime.

Note: Currently this method is only available through theservice endpoint. Ingesting through the Analytic UI does not usethis endpoint.

Expected Folder Structure

The following diagram represents the expected file and folder structure for an analytic ingestion .zip file.All folders and files indicated below are required for ingestion through the /analytics/ingestionsendpoint. The diagram indicates the required folders and files for ingestion through the /analyticEntries/ingestions endpoints in blue. It indicates required folders and files for ingestionthrough the /deployments/ingestions endpoint in teal.

Figure 4: Ingestion File Structure Diagram

106 © 2020 General Electric Company

Page 111: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Schema Definition for Analytic Ingestion Files

Each analytic .json file intended to be uploaded through the analytic ingestion services must conform tothe respective schemas represented in this document.

AnalyticDefinition.json

This file defines the ingestion metadata contract for AnalyticEntry. This file contains the metadatainformation for the analytic template such as name, URI, author, and contract version. It also has theanalytic template configuration and definitions including inputs, outputs, constants, and supporting fileswith relative paths to the analytic definition file. Application Analytics does not validate the file nameuniqueness and expects the user to keep track of file names. The analytic template metadata andconfigurations file, for example, analyticdefinition.json, must be placed inside the analyticEntriesfolder.

The following code snippet indicates the JSON structure with required (mandatory) field configurations.The file must conform to the schema indicated in the code snippet.

{ "@type": "analyticEntrySsRmMetaData", /* analyticEntrySsRmMetaData is for RM, analyticEntrySsMetaData for MEH , analyticEntrySparkMetaData for Spark*/ //Mandatory "contractVersion":"1.0.0", // Mandatory "name": "Analytic1", //Mandatory "uri": "/analytics/16a47853-0804-4972-ac33-b16d33336352", // Not always required. In case of Smart Signal, this is provided. CAF preserves it. Mandatory for SmartSignal "author": "GE Digital - Smart Signal Analytics Team", //Mandatory "description": "This is UC_Temperature_Asset's description:", "analyticRuntimeType": "SmartSignal", /* will be Smart Signal */ //Mandatory, Same for bith MEH and RM "primaryCategory": "Supervised Learning", //Mandator "secondaryCategory": "Supervised Learning", /* OPTIONAL */ "analyticVersion": "1.0.0", "hasAnalyticArtifact": false, //Mandatatory "analyticArtifact": { //Mandatory "language": "SmartSignal", "languageVersion": "v1" }, "executeParams": { // This is specific to Smart Signal and execute Params are not required. CAF uses this as a default for Smart Signal deployments. "repeatData": { "timeValue": 1, "timeUnit": "MINUTES" } }, "isReadyOnly": true // START Mandatory "analyticDefine": { "supportingFiles": [ { "filename": "../supportingFiles/blueprint.json", "description": "This is UC_Temperature_Asset's description:" }, { "filename": "../supportingFiles/modeblueprints.json", "description": "Mode Blueprints" },

© 2020 General Electric Company 107

Page 112: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

{ "filename": "../supportingFiles/modelblueprints.json", "description": "Model Blueprints" }, { "filename": "../supportingFiles/assetblueprinttags.json", "description": "Asset Blueprint Tags" }, { "filename": "../supportingFiles/modelblueprinttags.json", "description": "Model Blueprint Tags" }, { "filename": "../supportingFiles/assetblueprinttagrules.json", "description": "Asset Blueprint Tag Rules" }, { "filename": "../supportingFiles/modelblueprinttagrules.json", "description": "Model Blueprint Tag Rules" }, { "filename": "../supportingFiles/ruletagassociations.json", "description": "Rule Tag Associations" }, { "filename": "../supportingFiles/diagnostics.json", "description": "Diagnostic Rules" } ], "inputs": [ { "name": "temperature", "unit": "", "description": "temperature", "dataType": "Double" }, { "name": "pressure", "unit": "", "description": "pressure", "dataType": "Double" } ], "constants": [ { "value": 0.92, "uri": "", "name": "threshold", "unit": "", "description": "threshold", "dataType": "Double" } ], "outputs": [ { "name": "power", "unit": "", "description": "", "dataType": "Double" }, {

108 © 2020 General Electric Company

Page 113: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"name": "heatrate", "unit": "", "description": "", "dataType": "Double" }, { "name": "availability", "unit": "", "description": "", "dataType": "Double" } ] // END Mandatory "outputEvents" :[ "name": "<your alert template name. Must be present in the tenant. Validated. >", ] }}

deployments.json

This file defines the ingestion metadata contract for all deployments. Each deployment .json file mustbegin with the prefix deployment_. This file contains the following information required to configureanalytic deployments:

• Link to template artifact• Name for this deployment• Author of this deployment• Asset information - which asset this analytic is applied to• Link to IOMapping artifact files.

The following code snippet indicates the .json structure with required (mandatory) field configurations.The file must conform to the schema indicated in the code snippet.

{ @type: "deploymentSsRmMetadata", // Mandatory "contractVersion": "1.0.0", //Mandatory "name":"<your deployment name>", //Mandatory "templateType": "Analytic", /* Analytic or Orchestration */ //Mandatory "templateName": "<your template name from the template.json ingested" //Mandatory "templateVersion": "<version>", "isReadOnly":true, //Mandatory - If set to false, the deployment configuratons can be modified in the Analytic UI. "keyValueStore" :{ "runtimeAnalyticUri":"<your url CAF will call to get custom data for deployment>", //Mandatory for RM Smart Signal Ex: /analyticIntance/13a47853-0804-4972-ac33-b16d33336351) targetCardType:"SmartSignal", //Mandatory "<key1>":"<value>", ... }, "entityFilterFileName": "../entityFilter.json", //Mandatory, this is the relative path of the entityFilter.json "deploymentSteps": [ {

© 2020 General Electric Company 109

Page 114: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"analyticName": "<name>", //Mandatory for Template Type "Orchestration" "analyticVersion": "<analyticVersion>", "ioMappingFiles": [ "../ioMappings/IOMappingMetadata1.json", "../ioMappings/IOMappingMetadata2.json" // Multiple only if there is iterations. ] }] "deployParams": { /* Scheduling and data parameter information can be given here, This is not required in case of RM analytics */ "analyticArtifactVersion": null, "repeatData": null, "startTime": "1492723560", "endTime": "1493069160", "samplingInterval": null, "sampleDuration": null, "startOffset": null, "deployType": "DeployOnDemand", "doExecuteJob":false, "doDeployPerAsset": true }, "doDeploy": true}

ioMapping.json

Contains the mapping information for inputs and outputs in the template to asset tags.

The following code snippet indicates the .json structure with required (mandatory) field configurations.The file must conform to the schema indicated in the code snippet.

{ "@type": "ioMappingSsRmMetaData", // Mandatory "contractVersion": "1.0.0", //Mandatory "name": "ioMappingSsRm1", // Mandatory "description": "description", "inputsToDataSource": [ //Mandatory Start here { "portName": "EXHAUST_STEAM_TEMP_1", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "EXHAUST_STEAM_TEMP_1" }, "assetSourceKeyToDataSources": {} }, { "portName": "GROSS_POWER", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "GROSS_POWER" }, "assetSourceKeyToDataSources": {} }, { "portName": "INLET_STEAM_TEMP_1",

110 © 2020 General Electric Company

Page 115: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

"commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "INLET_STEAM_TEMP_1" }, "assetSourceKeyToDataSources": {} }, { "portName": "MODE_TAG_1", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "MODE_TAG_1" }, "assetSourceKeyToDataSources": {} }, { "portName": "INLET_STEAM_PRESS_1", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "INLET_STEAM_PRESS_1" }, "assetSourceKeyToDataSources": {} }, { "portName": "EXHAUST_STEAM_PRESS_1", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "EXHAUST_STEAM_PRESS_1" }, "assetSourceKeyToDataSources": {} } ], "constantsToDataSource": [ { "portName": "TURBINE_EXHAUST_STEAM_QUALITY", "commonDataSource": { "@type": "ioDataSourceValueCommon", "dataType": "Integer", "value": 0.92 }, "assetSourceKeyToDataSources": {} }, { "portName": "MODE_TAG_1_THRESHOLD", "commonDataSource": { "@type": "ioDataSourceValueCommon", "dataType": "Double", "value": 30.0 } } ], "outputsToDataSource": [ { "portName": "MODE_CV", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "MODE_CV"

© 2020 General Electric Company 111

Page 116: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

} }, { "portName": "TURBINE_PRESS_RATIO_CV", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "TURBINE_PRESS_RATIO_CV" } }, { "portName": "TURBINE_EXHAUST_STEAM_QUALITY_CV", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "TURBINE_EXHAUST_STEAM_QUALITY_CV" } }, { "portName": "TURBINE_EFFICIENCY_CV", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "TURBINE_EFFICIENCY_CV" } }, { "portName": "TURBINE_TEMP_RATIO_CV", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "TURBINE_TEMP_RATIO_CV" } }, { "portName": "CVMEHtag", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "CVMEHtag" } }, { "portName": "CVMEhTagF", "commonDataSource": { "@type": "ioDataSourceTagCommon", "entityType": "Asset", "tagName": "CVMEhTagF" } } ] //Mandatory Ends here}

entityFilter.json

This entity filter is created and is referenced in the deployment as the asset selection for the deployment.

112 © 2020 General Electric Company

Page 117: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

The following code snippet indicates the .json structure with required (mandatory) field configurations.The file must conform to the schema indicated in the code snippet.

{ "contractVersion": "1.0.0", "entityFilter": { "name": "XE7FA_297479_3504_Ingestion", "entityFilterConditions": [ { "entityFilterFieldName": "Asset Name", "operator": "=", "value": "XE7FA_297479_3504" } ] }}

© 2020 General Electric Company 113

Page 118: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Chapter

10Release NotesTopics:

• Fourth Quarter of 2019• Second Quarter of 2019• First Quarter of 2019• Fourth Quarter of 2018• Second Quarter of 2018• First Quarter of 2018

114 © 2020 General Electric Company

Page 119: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Fourth Quarter of 2019

AnalyticsThis topic provides a list of product changes released for this module on the specified dates.

Release Date: December 13, 2019

Table 22: Enhancements and New Features

The following enhancements and new features have been added.

Description Tracking ID

The following are the Orchestration Scheduling Enhancements:

• Data request parameters can be set on per step basis. Data request parameters, such as

Start Time, End Time, for Run Once, Offset before schedule, and Sample duration, which are

set at orchestration level, apply to all steps by default. However, you can set the parameters

on per step basis and override these parameters.

• In addition to the Historian mode, the following modes are available in the Time Series data:

◦ Math

◦ Sample and Hold

• Runtime supported for Predix Insights only.

US363606

Table 23: Obsolete Features

As of this release, the following items are no longer available.

Description Tracking ID

CAF no longer supports deployment of analytics written in Python 2.X. The user interface change

for CAF that disables creating analytics with Python 2.7 will be introduced in APM 2019 Q4

release. If you have a legacy code and must use an earlier version, you must make sure that the

library versions in the config file point to the earlier versions. Python uses the new version by

default, which may not be backward compatible. This may work if there are minimum conflicts

from a single library.

Note: Existing analytics that are deployed to PAF will continue to work, but new deployments

need to use Python’s newer version 3.X when creating an analytic in CAF.

US363606

© 2020 General Electric Company 115

Page 120: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Second Quarter of 2019

AnalyticsThis topic provides a list of product changes released for this module on the dates listed below.

Release Date: June 28, 2019

Table 24: Enhancements and New Features

The following enhancements and new features have been added.

Description Tracking ID

Using the common analytics framework, you can now generate multiple alerts at the same time

to analytics that are deployed on the Predix Insights and Predix Framework runtime

environments. You can also associate one or more alert templates with an analytic using the

Output Events section.

Note: To associate an alert template with an analytic and to generate multiple alerts, you must

enable the template in the I/O Mapping page.

US355437

Release Date: June 7, 2019

Table 25: Resolved Issues

The following issues that existed in one or more previous versions have been resolved.

Description Tracking ID

Previously, you could not access Asset Filters through the module navigation menu if you

selected a navigation profile from the Module Navigation section under Tenant Preferences.

This issue has been resolved.

DE109144

First Quarter of 2019

AnalyticsThis topic provides a list of product changes released for this module on the dates listed below.

Release Date: March 29, 2019

Table 26: Enhancements and New Features

The following enhancements and new features have been added.

116 © 2020 General Electric Company

Page 121: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Description Tracking ID

To enhance the usability in analytic template and deployments, you can now download a .zip file

of logs containing all analytic deployments for the current version of the Analytic Template.

F42079

To enhance the usability in analytic deployments, you can now configure your tenant to connect

to an instance of PostgreSQL that has been previously configured on Spark runtime as one of the

input data source.

• F42077

• F42078

To enhance the usability in analytic orchestrations, you can now schedule analytic orchestrations

to run in a recurring mode for Spark runtime.

F42076

To enhance the asset applicability during deployment, the asset applicability can now be derived

from existing deployments and asset filters associated with each deployment. Newly

commissioned assets that fit the filter criterion are automatically available for selection during

deployment.

F42074

To enhance the I/O Mapping step during deployment, you can now configure the Entity Type for

inputs, outputs, and constants in the analytics template. During the I/O mapping step in

deployment, the Entity Type tells Application Analytics to retrieve the tags and attributes of all

assets and their children starting at the defined level.

Note: This is not available for all runtimes.

F42072

To enhance the usability in the analytic template creation, you can now associate a Machine

Learning (ML)/Artificial Intelligence (AI) model to an analytic or blue print. This is only available

for the PowerDataFabric runtime.

F42071

To enhance the Scheduling step during deployment, you can now trigger an analytic to run when

a specific alert event happens.

Note: This is not available for all runtimes.

F18922

Table 27: Resolved Issues

The following issues that existed in one or more previous versions have been resolved.

Description Tracking ID

Previously, when a deployment used more than 1000 assets but did not have all outputs explicitlymapped, then the error ConcurrentModificationException occurred and thedeployment failed. This issue has been resolved.

DE97332

Previously, on dedicated tenants, administrators could not access the audit logs for Analyticsusing the Audit Log Management. This issue has been resolved.

DE94501

Previously, sometimes, even though there were no current deployments associated with theversion of the analytic templates you were trying to delete, an error occurred asking you to firstdelete any associated deployments before you can delete the version of the analytic artifact inthe template. This issue has been resolved. You can now delete analytic templates which have nocurrent deployments.

DE77551

Previously, when you modified an Analytic Template configuration after it had already been usedin a deployment, the deployment status in the Deployments page was expected to displayUpdates Not Deployed. Sometimes, the status did not automatically refresh. Thisissue has been resolved.

DE63501

© 2020 General Electric Company 117

Page 122: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 28: Known Issues

Description Tracking ID

If a deployment uses more than 250 assets that generate alerts at frequent intervals of less thanone hour, you may not be able to access these alerts. There are some known limitations whenusing a combination of high-volume and high-frequency alerts generated by analytics.

If you have a need to deploy analytics that are expected to produce alerts at high frequencies andfor large volume of assets (over 250), please contact the Analytics Support Team prior todeploying such analytics.

This limitation is limited to the Predix runtime environment.

A future enhancement for rendering high-frequency, high-volume analytics alerts is scheduled fora future release. At that time, all existing deployments must be re-deployed to benefit from theenhancement.

DE105943

Limitations

• Application Analytics does not support cross-tenancy; therefore, you will not be able to access assetsfrom a different tenant.

Fourth Quarter of 2018

AnalyticsThis topic provides a list of product changes released for this module on the dates listed below.

Release Date: December 14, 2018

Table 29: Enhancements and New Features

Description Tracking ID

To enhance the performance during orchestration deployment, Application Analytics now caches

data between orchestration steps.

F37992

To enhance the usability in accessing relevant deployments, you can now:

• Filter the deployment results using the deployment name or Analytic Template name in the

keyword search.

• Search and retrieve related deployments by assets or entity ID. If the asset ID matches one

or more of the filter criteria defined in an asset filter that was used in deployments, then all

related deployments are listed for that asset ID.

F36287

To enhance the modifications to orchestration deployments, you can now define overrides for

asset applicability at the input level during deployment.

Note: This is not available for all runtimes.

F36279

To enhance the usability in the deployment creation and modification process, in the I/O Mapping

step during deployment, you can now expand and collapse the ASSET(S) and Data Flow panes.

Additionally, you can now:

• Access tags and attributes at the asset level.

• Map tags and attributes that are not common to all assets.

• Access the list of steps and iterations as a tree view in the Data Flow pane.

• F36275

• F36274

118 © 2020 General Electric Company

Page 123: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Description Tracking ID

To enhance the I/O Mapping step during deployment, you can now configure an Analytic

Template to allow mapping inputs to aggregate tags or multiple assets. This also provides the

ability to write back to tags defined at any asset hierarchy level.

Note: This feature is limited to the Spark runtime.

• F36259

• F18099

To enhance the asset applicability during deployment, you can now configure asset filters with

Custom Attributes from non-asset entities.

Note: This enhancement is limited to AIR and SOAR runtimes only.

F33365

To enhance the retrieval of inputs and storage of outputs during deployments, you can now

connect to the following data stores:

• GE Digital APM Time Series

• GE Digital APM Asset Attributes

• Blob Store

F26075

To enhance the usability, you can now define data points for analytic inputs during analytic or

orchestration deployment.

F26069

To enhance the ability to configure the Analytic Template, you can now make constants optional

in the Analytic Template definition.

F17879

To enhance configuration in the Analytic Template, you can now define the unit of measure

conversion for an analytic input or output. You can then pass that conversion formula to the

respective runtime during deployment.

F10674

To enhance the ability to assign and control Analytics access permissions, you can now define

permission sets using multiple features and permissions instead of one large permission set.

F2568

To enhance deployment modifications, you can now modify the version of the Analytic Template

in deployment. If the Analytic Template has multiple versions of the analytic artifact uploaded to

the catalog, the user can select which version to apply during deployment.

US231988

To enhance usability in the Analytic Template creation process, you can now see a customized

form with only relevant fields and boxes specific to the selected runtime. Other options not

relevant to the runtime are hidden.

US112770

Table 30: Resolved Issues

Description Tracking ID

Previously, when using the .csv file, the Format Value column did not appear and you could onlydefine values for the Mapping Value Type field in the Analytics Templates page. This issue hasbeen resolved. Now, you can use the .csv upload template to define values for the Data Formatfield as expected.

DE78361

Previously, after tenant creation, the default asset filters and filter field data were unavailable forthe new tenant that was set up in QA2 environment. This issue has been resolved.

DE63501

© 2020 General Electric Company 119

Page 124: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Description Tracking ID

Previously, an Analytic Template failed to deploy or run if the Sampling Interval for interpolateddata was set to 0 minutes or 0 seconds even though the deployment was saved. Thisissue has been resolved.

DE62748

Previously, for analytics that were deployed and executed in the Predix Analytic runtime, if youredeployed the analytics with a newer version of that Analytic Template that had been uploadedto the catalog, the deployment failed, the deployment status was set to Not Deployed,and the following error occurred: invalid service task name <task id>.This issue has been resolved. Now, if you redeploy a previously executed analytic with a newerversion of the Analytic Template that had been uploaded to the catalog, the Deploymentcompletes as expected.

DE62740

Table 31: Known Issues

Description Tracking ID

If a deployment uses more than 1000 assets but does not have all outputs explicitly mapped,then the error ConcurrentModificationException occurs and thedeployment fails. If you notice this error, make sure to map all outputs before scheduling thedeployment.

DE97332

If you make modifications to an Analytic Template configuration after it has already been used ina deployment, the deployment status in the Deployments page is expected to displayUpdates Not Deployed. Sometimes, the status does not automatically refresh. If younotice this behavior, simply refresh the page for the status to display correctly.

DE95790

Only on dedicated tenants, administrators cannot access the audit logs for Analytics using theAudit Log Management page. Administrators should contact support to access these logs viathe backend. This is not an issue on shared tenants.

DE94501

Sometimes, even though there are no current deployments associated with the version of theanalytic templates that you want to delete, an error occurs indicating that you must first deleteany associated deployments before you can delete the version of the analytic artifact in thetemplate.

DE77551

Limitations

• Application Analytics does not support cross-tenancy; therefore, you will not be able to access assetsfrom a different tenant.

120 © 2020 General Electric Company

Page 125: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Second Quarter of 2018

AnalyticsThis topic provides a list of product changes released for this module on the dates listed below.

Release Date: June 29, 2018

Table 32: Enhancements and New Features

Note Tracking ID

Added support for flexible analytic input mapping, allowing multiple analytic inputs to be mapped

to the same input tag.

F34828

Deploy Analytics to assets in an equipment template.

• During the 1.Asset Selection step in Deployment, search and browse through a list of

available equipment templates to select an equipment template for deployment.

• Use Tags and Attributes available in the Equipment Template to perform I/O mapping for the

analytic inputs, constants, and outputs.

F33386

Access the SmartSignal Analytic Maintenance application from the Deployment page in

Application Analytic app.

F34831

Added support to schedule analytics using raw data per input tag. The Tabular and Interpolated

data options are still supported.

F34847

Ability to map asset attributes to analytic input and output definitions, so that the analytic inputs

can be retrieved and results written to the configured asset attribute.

F30566

Ability to provide notes and disposition in Alerts ingestion JSON. F34846

Table 33: Known Issues

Note Tracking ID

The CSV file upload is one way to quickly import input definition, constants, and output

definitions into an existing analytic template. However, the CSV template does not provide the

column to input the Data Format values. You must define the Data Format only through the UI.

The updated CSV template will capability will be available in the next release.

DE78361

Release Date: April 27, 2018

Table 34: Enhancements and New Features

Note Tracking ID

Ability to ingest, configure, and deploy analytics to the Edge Analytics Runtime (Limited

Availability - available on request only)

• Users can download and install the Edge Analytics Editor to create analytics. This is

accessible from the Edge Manager repository for the (cloud) application tenant.

N/A

© 2020 General Electric Company 121

Page 126: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Table 35: Resolved Issues

Note Tracking ID

For analytics deployed to the Spark Runtime and scheduled to Runonce; even though the

interpolated option is selected, the data always returns as raw data from timeseries

instead of the expected interpolated results.

DE69377

Even though a deployment only requires either the output definition or output event to be

configured correctly in an analytic template, Application Analytics does not allow the user to

create deployments using a template that does not have output definitions. Even if the

Deployment tab is enabled, the user cannot create deployments and the following error

message appears: Unable to create deployment as there are no'outputs' for the <analytic_template> analytic.

DE64793

Table 36: Limitations

Note Tracking ID

See the March 30, 2018 release notes for Edge Analytics. N/A

Table 37: Known Issues

Note Tracking ID

See the March 30, 2018 release notes for Edge Analytics. N/A

First Quarter of 2018

Analytics Release Notes for February 8, 2018Provides a list of new and updated features for this release, as well as any resolved issues, limitations, andknown issues.

Release Information

Module Name Analytics

Release Date February 8, 2018

Documentation Build Date February 8, 2018

What's New or Changed in This Release

• Bring Your Own Runtime - Ability to register, enable for a tenant, and utilize an external customruntime with GE Digital APM Application Analytics. This feature is release with limited availability.

• Ordering of deployments in Orchestrations

◦ Ability to assign a priority value to individual deployments in the 3. Schedule page.◦ This is enabled only for deployments associated with analytics orchestrations for a registered,

external custom runtime.

122 © 2020 General Electric Company

Page 127: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Limitations and Known Issues

Tracking ID Description

DE62740 For analytics that are deployed and run in the Predix Analytic

runtime, redeploying with a newer version of the analytic

uploaded to the catalog results in a failed deployment with the

newer version. The deployment status resets to NotDeployed and the system responds with an error

invalid service task name <task id>.

Workaround: For all analytics that must be deployed to the

Predix Analytic runtime, make sure that you create a new

deployment for every time a new version of analytics is

uploaded to the catalog. All existing deployments will continue

to use the older version of the analytics used to first deploy it.

DE62748 The Analytics Deployment will be saved; however, the Analytic

will fail to deploy or run if the Sampling Interval for

interpolated data is set to 0 minutes or 0 seconds.

The minimum expected value is 1 second.

DE63501 After tenant creation, default asset filters and filter fields data

may not be available for the new tenant that was set up in QA2

environment.

If this happens then, do the following:

Note: Only a system administrator can perform these steps.

• Sign in to your tenant with your administrator credentials.

• Navigate to the Permission Sets page.

• Access the AnalyticsExecutorRole default permission set

in the edit mode.

• If the option Manage Analytics Catalog is selected, then

clear the checkbox to remove the selection, and then select

Save to save the configuration.

• Navigate to the page again, and then select ManageAnalytics Catalog, and then select Save to save the

configuration again.

© 2020 General Electric Company 123

Page 128: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Resolved Issues

Tracking ID Description

DE56659 Intermittently, during analytics deployment, even after selecting

the asset in the 1. Asset Selection page, the Save, Reset, and

Next buttons may remain disabled.

Workaround: Navigate to the main Deployment tab, then select

the deployment. Select Edit to re-attempt the 1. AssetSelection and 2. I/O Mapping pages.

DE56646 After selecting Deployments in the main navigation, the status

of a deployment in the deployments list may not match the

status listed in the Deployments details page. For example, the

deployment list may show the status In Progress for

the deployment Deployment1, but the Details section for

the same deployment lists the status Completed.

Workaround: Refer to the deployment status on the details

page instead of status in the list.

DE56620 In Deployments, 2. I/O Mapping page, on attempting to clear

the value in the Data Points field for an analytic step or its

iteration, the system may default to an older, previously saved

value.

Workaround: At this time, configure the affected Data Pointsfield to a higher value than the number of points configured for

the Date Range in the 3. Schedule page; the system will then

update the Data Points value for the step or iteration.

DE56563 This defect only applies to analytics that use the

resolveAndExecute API for event-based triggering. If

analytics deployments were scheduled for event-based

triggering, and the 1. Asset Selection page was configured

using a default asset filter such as All Assets, All Enterprises,

All Segments, and All Sites, the deployments will not run

correctly.

Workaround: At this time for analytics that use the

resolveAndExecute API for event-based triggering,

create custom asset filters and use them to configure assets in

the 1. Asset Selection page in Deployments. Do not use

default asset filters to configure assets selection during

deployment.

124 © 2020 General Electric Company

Page 129: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Tracking ID Description

DE57027 In Deployments, 1. Asset Selection page, if you had previously

configured assets with a custom asset filter, and later updated

to select a different asset filter containing the same filter

criteria (for example, both asset filters produced the same

results of assets), the system may revert the filter to the older,

previously selected version.

Navigate to the main Deployments page, select the

deployment, and then select Edit to re-attempt asset selection

configuration. Alternatively, if the filter criteria are the same in

both filters, use the previously configured asset filter.

DE56906 In the Deployments details page and 2. I/O Mapping page,

analytics that have long names appear with the name and

version columns overlapping to obscure some text.

Workaround: Navigate to the Deployments details page. You

can note the name of the analytic orchestration heading in the

deployment header. You can then navigate to the

Orchestrations list, and then search for that specific

orchestration and access it in the details page. You can then

access the analytics template names and version used in the

data flow.

Analytics Release Notes for March 30, 2018Provides a list of new and updated features for this release, as well as any resolved issues, limitations, andknown issues.

Release Information

Module Name Edge Analytics

Release Date March 30, 2018

Documentation Build Date March 22, 2018

What's New or Changed in This Release

• Support for Analytics/Orchestrations on Spark Runtime (Limited Availability)

◦ Uses the Predix Insights (Beta) capabilities.◦ Orchestration data flows are supported for both Java and Python based analytics in this runtime.◦ Provides the ability to modify and delete orchestration deployments.◦ Provides the ability to map an existing output tag as the input for the next step in the orchestration

data flow. This functionality is limited to configuring already ingested tags in the asset model. It isnot possible to use output tags generated by the analytic during runtime.

• Ability to Ingest, Configure and Deploy analytics to the Edge Analytics Runtime (Limited Availability -available on request only)

◦ Users can download and install the Edge Analytics Editor to create analytics. This is accessible fromthe Edge Manager repository for the (cloud) application tenant.

© 2020 General Electric Company 125

Page 130: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Limitations

• Analytics/Orchestrations on Spark Runtime has the following limitations:

◦ Orchestration deployments for Spark runtime can only be scheduled using the RunOnce andStreaming options.

◦ Orchestration deployments for Spark runtime scheduled using Streaming option limits it to onestep only.

◦ Spark runtime does not cache the analytic outputs during analytic execution.◦ No fallback supported if an orchestration deployment fails in the Spark runtime after running the

deployment.• Edge Analytics runtime has the following limitations

◦ Currently Edge Analytics only supports OPC UA for automated data collection and deployment.◦ While most functionality is available, there is no support to modify existing Edge Analytics

deployment instances.◦ You must configure and set up the Edge Manager separately. For example, installation of Edge

Analytics runtime and other software must be performed separately.◦ Deployments are sequential: only a single Deploy or Delete operations can be done at any time.

You must first complete the action before another user initiates the next Deploy or Deleteoperation.

◦ Deployment Deletion

◦ Failed deployments, that go from, In Progress status to Not Deployed status, must be:

1. Deleted from Deployments list in Application Analytics2. All CAF_Configuration files and Analytics on each Edge Device must be manually deleted

(cleaned up) using Edge Manager to ensure that Edge Device state remains consistent withDeployments listed in Application Analytics.

◦ Application Analytics automatically deletes deployments in the Edge Devices, when you deletedeployments in Application Analytics that have successful Edge Analytics Deploymentsstatuses of Deployed, or Scheduled.

◦ While a Delete operation is in progress, indicated by spinning wheel - wait/busy icon, a usermust not navigate away, but just wait until busy icon disappear. The deployment is removedfrom Deployments, which indicates that Delete operation has been completed. After which, theuser should confirm in Edge Manager that all analytics, and related configuration to thisDeployment has been removed/updated on the relevant Edge Devices.

◦ After a deployment has been deleted, you cannot reuse the same name for another EdgeAnalytics Deployment.

126 © 2020 General Electric Company

Page 131: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Known Issues

Tracking ID Description

DE64497 For analytics deployments to Edge Analytics runtime, during the

2. I/O Mapping step in Deployment, the Deploy button

remains disabled if optional input fields are not fully mapped

even though it is not required.

Make sure that during 2. I/O Mapping step in Deployment, all

input tags are fully mapped before proceeding irrespective of

whether or not they are required inputs.

DE67658 In Predix runtime, tags conforming to the expected tag

nomenclature do not get listed as expected in I/O Mapping

when selecting the Map for All Assets and Add SuggestedTags options.

Workaround: Manually perform the I/O mapping to the correct

input and output tags.

DE67897 The Delete button remains disabled for deployments with the

status In-Progress for analytics deployed to the Edge

Analytics runtime.

Typically, even though deletion of deployments is not currently

supported for Edge Analytics runtime, the Delete button is

expected to remain enabled. When you select the button, an

error message In-progress deploymentscannot be deleted should appear.

DE64793 Even though a deployment only requires either the output

definition or output event to be configured correctly in an

analytic template, Application Analytics does not allow the user

to create deployments using a template that does not have

output definitions. Even if the Deployment tab is enabled, the

user cannot create deployments and the error message:

Unable to create deployment asthere are no 'outputs' for the<analytic_template> analytic appears.

© 2020 General Electric Company 127

Page 132: Analytics › ... › predix-apm › PDFs › Analytics.pdfSmartSignal, Predix, and Edge Analytics Runtime. The application provides services that can be extended to perform the following

Tracking ID Description

DE62724 For analytics deployed to SmartSignal runtime, in the rmdeployment- IOMappingMetadata1.json,

Application Analytics does not display a warning or error

message when the configured constant value does not match

the constant data type. This can potentially lead to deployment

failures.

Workaround: Always ensure that the constant values entered

must match the data type defined in the analytic template

configuration.

DE69377 For analytics deployed to the Spark runtime and scheduled to

Runonce; even though the interpolated option is

selected, the data always returns as raw data from timeseries

instead of the expected interpolated results.

Resolved Issues

Tracking ID Description

DE63501 After tenant creation, default asset filters and filter fields data

may not be available for the new tenant that was set up in QA2

environment.

If this happens then, do the following:

Note: Only a system administrator can perform these steps.

• Sign in to your tenant with your administrator credentials.

• Navigate to the Permission Sets page.

• Access the AnalyticsExecutorRole default permission set

in the edit mode.

• If the option Manage Analytics Catalog is selected, then

clear the checkbox to remove the selection, and then select

Save to save the configuration.

• Navigate to the page again, and then select ManageAnalytics Catalog, and then select Save to save the

configuration again.

128 © 2020 General Electric Company