This is little strange to have something like this.
What I remember, we used to have,
Suggested actions:
Raise an empower ticket stating the issue with screenshots.
Estimate the load in production and simultaneous thread required to run the load. [Try to find GEAR doc on server performance tuning. It used to have an excel sheet which had different sections, and tell you the size of your JVM heap]
What is your environment configuration look like? Take help of an expert to set different parameters to set the resource settings. [No straight forward answer, IMHO]. Heavy pub-sub, webservices, adapter notifications, etc etc.
Check and explore clustering advantages, if not available right now. If available check the other node(s) of cluster. Check your load balancing parameters.
Thanks for the quick checklist:
Let me give you a background of the env but before that : I think my issue is resolved with simple simple steps already
We had used the GEAR Template for capacity planning, but as you would be knowing it was created for older versions and for latest versions instead of GEAR there is another paid service.
We already had production environment on 7.1 version actually and we took similar parameters for resources, settings etc.
I had configures and changed the JVM settings during the time of installations and increased them.
However, I think I had missed the step that after changing it, we need to update installSvc.bat file in IS\support\win32.
I just did the same and restarted the IS and surprisingly am seeing more free memory and available threads withing the maximum limit…
M sure there is no surprise to it, perhaps this was to be done along with JVM setting changes.
In all 4 IS, now the parameters and readings are controlled actually.
May be if you had similar exp, pls share some of it, so that I could under stand deeply.
PS: I guess I shall observe for a few days and then raise aticket if reqd.
We are facing the same issue in our DTQ & prod environment, Raised an ticket with SAG & got the below response.
SAG Response
Are you refering to available threads showing more than 100% ?
If yes, it is a calculation issue rather than thread problem. This has
already been reported and will be fixed in IS_9.5_SP1_Core_Fix3 scheduled
to be released by end of this month
Bhanu: Thanks! looks like we too have to wait for Fix3. Yesterday, I did see that all parameters were conttrolled but again this morning we have the same issue recurred.
RMG: Yes, we had modified jvm heap : max to 4096 MB and min to 1024 MB for one of the IS.
We had this verified from prev verison and also from new template before changing it.