INC-169055 · Issue 646670
Performance improvements for database queries and archiving
Resolved in Pega Version 8.3.6
In order to improve the performance of database queries, repetitive unnecessary database calls have been removed, archival queries have been simplified, and diagnostic metrics were added. Optional batch requestors have been added to all archival services with their own DSS settings to allow processing more cases in parallel. It is recommended to create 2 additional indexes to help with the archival process. The indexes will be on:pegadata.pr_metadata (pyisparent, pyobjclass, pyarchivestatus) INCLUDE (pyinskey); pegadata.pr_metadata (pyisparent, pyarchivestatus) INCLUDE (pyinskey);
SR-D91660 · Issue 555722
Updated document ID handling for deleting indexes
Resolved in Pega Version 8.3.6
When trying to delete a document for index name "work*", intermittent errors were seen indicating an invalid index name or index does not exist value. This has been resolved by updating the handling for delete by ID.
INC-155276 · Issue 622816
Null check added for step page
Resolved in Pega Version 8.3.6
After creating and adding new Access Roles and application 'Access When' to the privileges instead of Production level, during run time the error "runtime.IndeterminateConditionalException: Trying to evaluate Rule-Access-When conditions L:IsProdAccess when there is no page to evaluate them against" appeared for the specific privileges. This was traced to a missed use case where the system falls back to the step page if the page for evaluating the 'when' condition is null, which did not account for scenarios where the step page can be null. To resolve this, a null check has been added which will fetch the primary page if the step page for the access 'when' condition is null.
INC-156647 · Issue 626293
Improved disconnected requestor cleanup for FieldService
Resolved in Pega Version 8.3.6
A large number of requestors from FieldService with the status as 'Disconnected' were accumulating and causing performance issues. This was traced to the requestors not getting passivated due to users not logging out and new requestors being created for the same users next time, and was caused by the value of the DSS Initialization/PersistRequestor being set as "OnTimeout". When the DSS prconfig/timeout/browser/default is not configured, the default browser requestor timeout is 60 minutes. In this scenario, requestors were not passivating as the requestor passivation timeout was set to the refresh token lifetime for mobile users, which was very large and overwrote the DSS value. This has been resolved by removing the code which set the passivation timeout to the OAuth2 refresh token lifetime.