If there is no BLOCKED threads, then look for RUNNABLE threads but that should be runnable for long time.
For Eg : We have started OpManager two days before. Now we are taking Threaddump and we could see the DATAPOLL thread in runnable for last two days. If it so, That could be the problematic device which blocks the polling of other monitors as well.
If the RUNNABLE threads(waiting for long time), then try to address the issue with the problematic device
If the RUNNABLE threads are from different device and it is running for long time, then there might be some issue with those devices. Possible Causes seen earlier :-
SNMP thread get blocked while trying to find the DNS name of the device from SNMP API
We can easily find out with the above details.
If that is not the above case
Connect the database or Connect in SubmitQuery.do and do execute the below query
select period, protocol, count(*) from polleddata group by period, protocol order by period
Based on the count of polleddata, we need to tune the threads.conf file as well as wrapper.conf file with maxmemory settings
If there is more polleddata with minimum(<=60 sec) poll interval, check for the polls per second under Support page
If the polling interval is very minimum, please ask him to increase the polling interval as well
When working on increasing the threads and maxmemory in wrapper.conf file, then you have to look for the below things :-
RAM size of the server running OpManager
Is there any other running in the same server. If it so, we do not proceed with memory tunings as well.