Child case queries can be slow after upgrade
Valid from Pega Version 7.2.1
After upgrading, performance might be slow when you run queries on child cases for work tables that existed prior to the upgrade. This issue is often indicated by PEGA0005 (query time exceeded limit) alerts that can be correlated to these queries in the logs. To resolve the issue, add a database index to each of your work tables by using a multicolumn index on the pxCoverInsKey, pxObjClass, and pyStatusWork columns.
For more information, see Performance improvements for child case queries.
Enhanced entity extraction with the CRF method
Valid from Pega Version 7.2.1
The Free Text Model rule now uses the Conditional Random Field (CRF) method for named entity extraction. The CRF method improves the accuracy of entity extraction by using sequence tagging and by reducing the number of false positives that were generated in the process of extracting named entities from the text. In addition, you can define your own custom entities and import them into the Pega 7 Platform as entity extraction rules that contain Rule-based Text Annotation (Ruta) scripts.
For more information, see Improved Free Text Model rule type.
Support for Twitter data recovery
Valid from Pega Version 7.2.1
You can configure a Twitter data set to use the new playback option to recover tweets that the system could not retrieve during a period of time in which tweets were unavailable, such as during a data flow failure. You can specify the time period for which you want to retrieve Twitter data and enter the maximum number of tweets that you want to retrieve.
For more information, see Improved Free Text Model rule type.
Verify ML Readiness error: Upgrades from Pega 7.1.0 through 7.1.8 do not replace nonversioned rules
Valid from Pega Version 7.2
Upgrading from any Pega 7 Platform version between Pega 7.1 and Pega 7.1.8 to Pega 7.2 does not automatically replace nonversioned rules. If you revalidated and saved a Pega 7.1.x rule after the timestamp on the Pega 7.2 version of the rule, upgrading to Pega 7.2 does not overwrite the saved rule. You might see either of the following behaviors:
- The upgrade fails with the following error message:
Verify ML Readiness - The upgrade succeeds, but you might be unable to open changed rule forms.
If the upgrade fails, see the Pega 7 Platform Upgrade Guide to force the upgrade.
This known issue does not affect Pega 7.1.9. After the upgrade has finished, apply Hotfix 27288. For information about obtaining and applying hotfixes, see Hotfix Manager.
Time-based retrieval of Facebook posts
Valid from Pega Version 7.2.1
You can now control the amount of data that is retrieved from Facebook data sets by using the new search functionality for Facebook data set records. You can limit the retrieval of posts to any number of past weeks, days, hours, or minutes to get only the data that is relevant to your business objective or that you lost as a result of a system outage or failure.
For more information, see: Improved Free Text Model rule type.
Improved monitoring of data flow runs
Valid from Pega Version 7.2.1
To monitor the status, progress, and statistics of a data flow run, open it on the Data Flows landing page. Use the run and distribution details to report on data flow progress and possible service-level agreement (SLA) breaches. Run details are available at the node and partition level. Component-level statistics are available for the run, nodes, and partitions.
For more information, see the Data Flows landing page.
Adaptive learning without using an interaction ID
Valid from Pega Version 7.2.1
When you configure the Strategy shape in a data flow to capture the response to previous decisions, you do not have to provide an interaction ID. Instead, during the process of adaptive learning, adaptive models can query the Event Store data set over a specified period of time to retrieve decisions that were made for each response. This way, you eliminate the need to store the interaction ID on the client side.
For more information, see Delayed learning of adaptive models.
Improved engine startup times
Valid from Pega Version 7.2.1
Performance enhancements in this release decrease engine startup times for new Java virtual machines (JVMs) and improve rule assembly for libraries. JVM startup times are faster because rule utility libraries and functions are no longer generated at startup. For more information, see Extract marker file not created at startup.
Single point of entry for decision management services configuration
Valid from Pega Version 7.2.1
Use the Services landing page to configure and monitor the Decision Data Store, Adaptive Decision Manager, Data Flow, and Visual Business Director services. This method simplifies decision management configuration because you can manage these services in one place.
For more information, see Services landing page.
Adaptive Decision Manager service in the Pega 7 Platform
Valid from Pega Version 7.2.1
The Adaptive Decision Management (ADM) service is now native to the Pega 7 Platform and is supported by the Decision data node infrastructure. Because of this enhancement, you no longer need to manage the ADM service as an external service. In addition, the ADM service does not require dedicated hardware or operating system environments.
For more information, see Services landing page.