This release includes hotfixes to core workflow features:
- Workflow evaluation can now be based on subject qualifier expressions, making the application more responsive.
- Several workflow label evaluation issues have been fixed.
- A critical issue with many child workflow steps being created instead of a single step has been fixed.
- The failover monitor has been updated, allowing it to be configurable to work better with JVM garbage collection.
Contact Reach Engine support with any questions, and to ensure you are all set with the features you have become familiar with.
Reach Engine Core Updates
|Internal Tracking ID
Add polling logic to the runCommandStepExecutor
New logic has been added to QueryStep and SubmitHTTPStep to use the following attributes:
This allows for workflow authors to achieve a query, delay, and check type loops in a single workflow step.
This significantly decreases the amount of workflow steps, and database load.
Long GC causes unnecessary failover
|Failover time has been increased to avoid failovers, which reduces the number of errors that require user intervention.
/workflow/search is slow to load modals
|Overall performance for evaluating workflow search terms for selected assets has been improved.
Create a property to abandon scheduled workflows
|The cleanup service is now configurable, allowing workflows to be abandoned by the system automatically after a set duration.
Reach Engine Bug Fixes
|Internal Tracking ID
Duplicate workflow steps
Child workflow steps were being duplicated causing a drain on system resources.
This fix ensures only one instance of the child workflow step is executed.
Workflows listed in the modal are listed out of ABC order
|Workflows are now properly sorted in alphabetical order regardless of capitalization.
defaultDataExpression re-evaluating data definitions and overwriting its current value at the time of execution for the workflow header executionLabelExpression
defaultDataExpressions were re-evaluating data definitions and overwriting current values at execution time for the workflow header's executionLabelExpression.
This caused workflows to change data unexpectedly when evaluated as an expression within the workflow, affecting executionLabels, setContextData valueExpressions, and other data.
This caused the workflow to override data when a workflow attempts to update the data definition to another value instead of the set defaultDataExpression.
The benefit of this fix is that the execution labels are now correct, preventing data definitions from being overwritten.
Workflows cannot be abandoned in certain cases
|The step index setting for workflows that has been fixed so that steps are now in the correct order, allowing workflows to be abandoned as expected.
Failover logs are displayed even after the node is marked offline
|Additional failover log entries for unhealthy and deregistered nodes in the scheduler logs have been suppressed and will no longer appear.