i am new to the group and pretty much supporting WebSphere in general. I am being asked for dumps which i assume doing a kill -3 will get me the dumps but am questioning the trace part. is that setup through the Websphere application itself? i have been trying to read on it but without seeing it i am not sure. Another group manages that server and the actual admin console for Websphere.

the main issue and the reason why i ask is we had seen messages regarding threads in pending. i guess i was hoping to see what best practice was when these happen so i can nail down the root cause. in the logs i looked at it was complaining about .jsp files all day with null pointers then later in the day i see the logs start saying threads in pending and those were incrementing. would this cause the hung process in WAS?

kill -3 , typically have javacore ( called thread dump in Oracle/Sun jvm) 

you may add -Xdump:java+snap:events=user, restart JVMs , then kill -3 will have snap

another way: wsadmin  or WAS isc console 

for wsadmin: 

 1 . jvm1 = AdminControl.completeObjectName('type=JVM,process=server_name,*')

 2 . 

AdminControl.invoke(jvm1, 'dumpThreads')                                # thread 

AdminControl.invoke(jvm1, 'generateHeapDump')                      # Heap

AdminControl.invoke(JVM1,'generateSystemDump')                  # system core

I can't find snap trace generated by wsadmin from IBM wsadmin document , need other expert help

Thanks