Recently, i have observed that the Nimsoft probe ntservices (ntservices.exe) utilizing more CPU and causing critical utilization alerts.
Why does this happen, Have anyone faced this ?
If you have a solution, suggest me with the same.
Do you face this issue for all servers ? I don't see any abnormal from probe .
What version of the probe are you running?
I am current using ntservices version 3.24
Currently i am observing this issue on few Windows 2008 R2 64bit OS.
And i dont think it could be anything with specific to OS flavour, as we have few hundreds of servers running the same.
please check and see if you are running robot 7.80, if not try upgrading to that.
also check and see if their is any antivirus on that system scanning the nimsoft directory.
If that does not help you will probably need to contact support to have them help find the problem with the environment.
There used to be an issue with this prior to 3.22 version of the probe.
if you have not tried completed deleting the probe and reinstalling it that is another thing to try.
hope this helps
Have had support case open on just this topic since February. CA development (HCL) has provided test build that is under going testing to see if issue is resolved. CA was able to reproduce the behavior observed after period of time of the usual okie dokie dance with support. As part of testing it was confirmed down revision to v3.22 or earlier helped alleviate and problem was introduced in versions above v3.23 and above. At least this has been my experience and suggested workaround that CA support discovered during lab testing.
Edit: Other item discovered during troubleshooting. Probe logging was found rather lacking. It did not output status of enabled service profiles regardless of log level and only output the status of each service upon startup. CA took that feedback to enhance eventually to include the status of each service being queried every time. This would help locate actual bottlenecks or where probe maybe getting stuck executing query. Another item sought during input from development was probe query execution method. It was determined probe queries as asynchronous (single thread) which explains query duration as in comparison to scripted mock up using Powershell reading in CSV for service profile list with expected state.
This has been addressed. Please raise a case with CA for ntservices 3.24T2.