I am using IS and UM with JMS Configuration/triggers.
While analyzing some performance issues between IS and UM I came to notice that at some point if I enable some triggers( concurrent triggers with 10 threads) then I start to see that pub.jms:send perfomance takes up to 10secs executing. check attached image.
I have changed some configuration mainly adding/enabling “Producer Caching” at the JMS Connection Alias. Performance is better, but not still impact is considerable after enabling more triggers on the IS.
If I run with only 1 trigger enable it is performing of course withoud any problems.
Do you know what configurations can be improved either on IS or UM to get “pub.jms:send” more performance?
Which version of UM are you using. If it is UM 10.1, please install the latest fix. A performance issue which got introduced in one of the earlier fixes has been fixed.
Sorry, forgot to mention the version.
version 9.9 IS and UM
What is the volumes you are facing this issue with jms:send and can you also check the IS/UM resource traffic running that point of time assuming this is happening in pre or prod type environment?
This is PROD environment, for that topic at some point in time it gets 900 messages per minute.
once i start reading from the same topic with more threads then i see latency.
I must add:
Performance goes down while putting/sending to the topic - pub.jms.send.
But it is more evident when I enable in same IS node a multi-threaded trigger consumer reading from the same topic.