We have been seeing issues of slowness where when we analyze the logs (using PLA) we find that there was a long requesting user before the system crashed. What could be the possible reasons for getting a long requesting user error in the logs? The severity of the log was marked critical.
Basically the long requesting user can happen for many reasons, before your system gets crashed their might be any thread dump being generated, so i suggest you to analyze the thread dump to find the real root cause of the specific requestor which is waiting for the lock and what operation is trying to perform during that time.
First you need to identify on which activity or rule the requestor is keep on running, it would be better you can attach the log files for analysis. Please hide sensitive information before uploading the logs.