This release brings several enhancements to the installation of ThingWorx Analytics functionality. Two installers are now available, each in both native Linux and native Windows formats:

Property transforms can be used to automate simple data transformations just by configuring a new property. The time and complexity of writing services, events, or subscriptions is eliminated. Property transforms can also be used to prepare data from connected assets for predictive analytics or monitor the statistical transformation of a property for specific behavior so that an alert can be triggered. These transformations can also be used to build solution-specific visualizations (mashups or widgets) or to create alerts based on logged property values.

The new Property Transform microserver is installed as part of the Platform Analytics installer (see above). Its transformations can be accessed via your Thing model in ThingWorx Composer. The transformation types available include:

Preparing metadata for ThingWorx Analytics can be time consuming and complicated. In past releases, metadata was required to be in JSON format. To simplify metadata preparation tasks, a new DetectMetadata service has been added to the Data microserver. When the DetectMetadata service is called, it infers the metadata from the CSV data and auto-generates an AnalyticsDatasetMetadata infotable. You can review and modify the infotable before using it to create a dataset.

This functionality is available both to ThingWorx Analytics Server and in Analytics Builder.

In this release, a set of Dockerfiles is available to build Docker Images of ThingWorx Analytics. These images can then be run locally on a server or in a container orchestration platform such as Kubernetes. A Dockerfile zip provides all of the files and information necessary to build and manage Docker containers for ThingWorx Analytics components. The zip file includes binary software files, Dockerfiles, Docker Compose examples, and supporting documentation.

Anomaly Detection functionality is built into ThingWorx and is enabled by default. However, in previous releases, two external Anomaly Detection services were required to implement the functionality. Those services had to be downloaded, installed, and configured. In this release, Anomaly Detection has been enhanced to use the ThingWorx Analytics Server microservices. These services are installed via the Analytics Server installer (described in New Functionality above). This change streamlines the implementation of Anomaly Detection. No additional external services are required and in the ThingWorx Alert Processing Subsystem, ThingWorx only needs to be configured to point to an AnalyticsGateway.

If you want to migrate a set of existing anomaly models that were created in a previous release, a manual process is available. Keep in mind that the initial steps in the process must be completed before you upgrade ThingWorx or the ThingWorx Analytics Server. For more information, see Migrate Existing Anomaly Models in the Anomaly Detection section of the Analytics Help Center.

TA-134

Previously, no method was available to terminate long-running jobs in the ThingWorx Analytics Server. In this release, job termination has been updated so that any job, regardless of status, can be terminated and removed, including model training, profiles, signals, and data loading jobs. This expanded functionality is available in both the Analytics Server and Analytics Builder.

TA-1081

Previously, users could select and delete multiple jobs from the Analysis Jobs page in Analytics Manager. Now, you can set criteria to delete jobs by state, creation date, model, and provider.

AM-1228

While performing input mapping and result mapping of analysis events, you can select infotables on either the source or target side or both to map rows of the infotable.

AM-1301

AppKey and password storage has been updated to use KeyStore encryption. You need to encrypt the application key or user ID and password before you start the agent. For more information, see Build an Agent using Java SDK.

AM-1322

The MaxMessageSize property has been added to the config.properties file. This property specifies the maximum messages transfer limit (in bytes) between the agent and the server per request.

AM-1296

A new Export capability has been added to the Models list page in Analytics Builder. You can select a completed model, click Export, and download the model as a PMML file.

TA-2541

AppKey and password storage has been updated to use KeyStore encryption. Encryption occurs automatically during installation and a Security Management Tool is available to manage updates to existing encrypted passwords.

TA-2376

In this release, the integration between DataFlowML and ThingWorx Analytics has been updated to accommodate the 2.1 release of DataFlowML. These updates include changes to the way the custom processors are built into DataFlowML pipelines and the way they are configured for use.

TA-2451

The ThingWorx Analytics Server version display has been simplified on Analytics Builder Server Configuration page. Previously, a version number was displayed that encrypted both the ThingWorx release number and the Analytics Server version. Now, the version shows the ThingWorx Analytics Server release plus the build version. For example: 8.4.0-r5.

TA-1217

In this release, security was improved by closing several vulnerabilities.

TA-1289, TA-1727, TA-1729, TA-1741, TA-1742, TA-1743, TA-1744, TA-1745

Previously, if the ThingWorx Analytics Server shutdown, no helpful messaging was displayed in Analytics Builder. In this release, appropriate messaging has been added based on which page is displayed in Analytics Builder.

TA-1592

When Anomaly Alerts are configured using the ThingWorx Trial Edition, they fail to enter Monitoring state. They moved directly from Calibrating to Failed and an error is generated. This issue may continue in earlier releases, but for ThingWorx Analytics 8.4, the update to allow Anomaly Detection to use Analytics Server microservices has resolved the problem.

TA-1764

An error is generated in Analytics Manager when trying to upload a third-party model. The same model can be loaded successfully via services in ThingWorx Composer. The issue occurs because Analytics Manager assumes that every PMML model includes a transformation dictionary. In fact, a PMML model is valid without a transformation dictionary. The issue has been resolved by adding a check to determine if a transformation dictionary is included before proceeding.

TA-1851

TA-2154

When incomplete JSON metadata is provided, dataset jobs continue to run without failing and providing error messaging. Two enhancements in the 8.4 release have contributed to resolution of this issue:

TA-1946

When additional data is uploaded to an existing dataset, its field values are not available for selection when creating a filter on the dataset. This issue has been resolved by fixing the way cached values were being cleared.

TA-2107

When ThingWorx Analytics is upgraded, the ownership of a specific directory was not properly updated. As a result, when training jobs are run in Analytics Builder, error messages are generated. This issue has been resolved.

TA-2312

TA-2297

Ordinarily, the feature list in a signals job displays features arranged in descending order based on their predictive strength on the goal, as determined by mutual information scores. However, if the signals job is launched with the Redundancy Filter enabled, the features are ranked according to the amount of information gain each feature provides.

In the case of this issue, when the signals job is launched from within a predictive model, and the model is created with the Redundancy Filter enabled, the feature list is not ranked according to information gain and the Information Gain column is not populated. This issue has been resolved so that features in the signals job are ranked as expected.

TA-2394

In some cases, when more than 25 models have been created, job status for all model rows is displayed as Running, even though the associated training and validation jobs have already completed. The created models can be viewed and used for predictions, but the status displayed on the models list page is misleading. This issue has been resolved.

TA-2505

When running a request for Signals, and specifying maxAtATime = 2, Mutual Information (MI), in relation to the goal, is not returned for all field combinations as expected. Instead, individual (one-at-a-time) MI scores are returned for all fields and then are filtered down to the top 25% most relevant fields. Two-at-a-time signals are calculated only for those fields. There is currently no way to modify this filtering behavior.

TA-729

Analysis agents must be restarted after a ThingWorx server restart. Without restart of the agent, any new simulation jobs will fail to execute.

—

AM-1289

When a replay event is executed, the on-screen progress bar does not update as the replay event progresses. The issue is related only to the user interface (UI), and the actual execution of the replay is not affected. As a work around, refresh the browser or navigate to another tab and then return to the Replay tab. The progress bar will update. This UI issue will be resolved in a future release.

AM-1345

When filtering is applied to a list page with multiple pages (such as the Models list page), the filtered results are not aggregated for display on a single page. Instead, the filtered results remain on their original pages and multiple pages must be clicked through to find all the filtered results.

To improve handling of list page displays, a configurable parameter has been added to set the number of table items that can be displayed per page. A more permanent resolution for the filtering with pagination issue will be provided in a future release.

TA-1596

TA-1618

TA-1928

There are currently some limitations on the size and number of columns a dataset can contain. These limitations depend on the mix of data types your data includes. For specific information, see the Data section of Prepare Data and Metadata.

When trying to upload a dataset with a large number of columns, if the upload fails and an error message is generated, try the following workarounds:

TA-2599

When a new instance of ThingWorx Analytics Server is installed using the Linux installer, only the twas-analytics-worker services are configured to restart automatically when the server is rebooted. The other microservices can be started and stopped normally after installation, but they are not enabled for automatic restart after a server reboot. As a work around, run the following script to enable all services for automatic restart:

This issue will be resolved in a future release.

TA-2672

Non-administrator users are unable to view models, despite being included in the necessary user groups. They can view information on Data, Signals, Profiles, and Settings pages, but not Models. This issue will be resolved in a future release. Use the following workaround procedure to make models available to non-administrator users:

TA-2753

The View Datasets and View Models screens do not load properly in the IE 11 browser for the support Windows operating systems. This issue will be resolved in a future release. To avoid the issue, use another browser, such as Chrome, Firefox, or Safari.

TW-31953

When the Analytics Extension is deployed on the 8.3.1 version of ThingWorx platform, the display of some pop-up windows might look compressed in Analytics Builder and Analytics Manager. In some cases, scroll bars will allow you to view truncated fields. However, the best work around for the issue is to maximize your browser view so that the pop-up windows can display fully. This issue will be resolved in a future release.

TW-38342

ThingWatcher was discontinued as a standalone SDK in 8.3. However, ThingWatcher functionality is available in the Anomaly Detection features native to the ThingWorx platform. For more information, see Anomaly Detection.

ThingPredictor software media is no longer available for download as of 8.3. However, support will continue for 8.x users until the next major release. To replace ThingPredictor capabilities, an improved set of predictive services is being introduced in 8.3. For more information, see How Predictive Scoring Using the Analytics Server Connector Works.

DataConnect and the Data Analytics Definition are not available as of 8.3. Replacement functionality will be introduced in a future release.