We noticed the table PR_LOG_SYSTEMSTATE keep growing and filling up file system and slows down entire application. We do have System Cleaner agent running, but its not archiving/cleaning this table. We have this issue in only one region.
Looking at the logic of the SystemCleaner (step 18 for 7.2.2), it will purge the entries in the table older than 90 days. In your case, can you check there are entries older than 90 days? The timestamp used is pxCreateDateTime.
so the question is: how realistic is your load test? If yes, then we should look into this in more details (create a SR). Otherwise, just truncate it manually as the logic in the SystemCleaner agent only deletes entries older than 90 days. What is your pega version and how many nodes are running during the load tests?
Our pega version is 7.21, this is clustered environment, currently we running 2 nodes on this cluster. We do have other apps running on different jvms, but sharing same pega db. We dont have SMA setup and agent running on those nodes, should we setup SMA and run SystemCleaner on those nodes as well?