Over the last few months we have been having performance issues in Apps 9 that have been very difficult to narrow down to a specific cause. We have plenty of overhead in both the database and middleware servers so I don’t think it is a hardware or database performance issue.
What I have been able to determine is that when we get really severe performance issues that we have a large job queue (default - 0) of waiting jobs usually caused by a custom job taking longer than normal to process - say 20mins during which time 6 x 10 Invoice posting jobs enter the queue and wait to be processed, this then prevents other jobs from running and they get added to the queue compounding the problem (sometimes upto 900 in waiting state). During this time the system becomes unusable from logging on to simple queries.
So I guess I’m asking:
- Has anyone else experienced these type of issue?
- We have 10 batch queues all set to processes:1 and the majority of jobs are all in 0 - default. Should we have more queues and or allow more processes to run at once?
- In sys parameters we were logging completed jobs for 360 days. Would this have an impact?
- Is there another way to diagnose the performance issues? when this occurs IFS Mon and EE are both unusable