I am encountering an issue in Pega Constellation where the dashboard is rendering an incorrect case count.
The implementation follows Constellation design guidelines, and the underlying auto-generated Report Definition has “Remove Duplicates” enabled to address potential duplication caused by joins/associations. Despite this, the aggregated case count remains inaccurate.
Additionally, we observed that when all filter criteria are not explicitly selected, the result set becomes inconsistent, leading to discrepancies in the overall count. The configuration involves Insights leveraging associations, which may be introducing duplication at the data retrieval layer.
Has anyone faced similar behavior with Constellation dashboards or auto-generated reports?
Are there recommended approaches (e.g., query optimization, distinct aggregation strategies, or association handling) to ensure accurate case counts in such scenarios?
In my opinion, They tend to create duplicate rows, and even with “Remove Duplicates”, the count can still be inaccurate - especially when all filters are not applied.
What I would suggest is:
Avoid relying on the auto-generated Insight for counts
Create a custom Report Definition and use COUNT DISTINCT
Try to reduce joins or move that logic to data pages if possible
You can also go with RDB-List if the query is too complex and you need full control over SQL, but I’d keep that as a last option since it’s not very maintainable.
Is the insight on the Case / Data Object directly? Or a Data Page? This is an important distinction as the OOTB Insights, run on a queryable datapage that is specified as your default data page in the Sources tab of the object - HOWEVER, what runs is a virtual report definition, so that underlying report won’t actually control the results.
I would suggest raising to Pega Support so they can look at your configuration and advise on your local configuration, just in case there is a bug here.