Records Doesn't aggregated after rollup


I have an ingestion with data loading from a Kafka stream.
From some reason, I see rows in the output that contains the same values exactly, while I would expect those records to be aggregated.
After investigating further I found that the records were inserted from 2 different Kafka partitions. I saw that in the past (~2018) there was a similar bug for this scenario, but the bug should be already fixed.

Any idea what might cause this issue?



Silly question maybe, but when you look at the actual payload in the console for the tasks that are running, are the queryGranularity and rollup set as you would expect?