ou-of-memory on SUD
I have a java application that threw an out of memory on SUD incident. When looking at the memory dump, there was stil a large amount of memory available still. I was wondering what the conditions need to be for this alert to trigger or if someone could provide more insight?
This typically happens when the heap gets exhausted by "temporary" objects allocated and referenced by active transactions/threads. Once the OOM kicks in those threads/transactions get aborted and subsequently they no longer reference the objects that triggered the OOM.
The screen shot states that a GC was triggered, if that is done by dynaTrace, you should file a support ticket to change this. an OOM triggered Heap Dump should not trigger GC.
If this is done by the JVM, then we cannot change this.
A workaround for the second usecase is to define an incident based on either high heap memory or high number of GC invocations and trigger a trending dump (with GC turned off). that would give you an idea of the objects in memory and should be able to tell you the perpetrator.
Community Member of the Month
September 2018
Announcing Dynatrace's Community Member of the Month for September 2018, Muthu P.! Click here to read more! Employee Member of the Month
September 2018
Announcing Dynatrace's Employee Member of the Month for September 2018, Kris Z.! Click here to read more!