The Microsoft WinMgmt operating system (OS) process frequently pegs the CPU at 100 percent, and consequently affects all call processing at that time. The WinMgmt Process spikes on both the publisher and subscriber servers.
This issue is documented in Cisco bug ID CSCsf17895.
One possible workaround for this issue is to restart the server or kill winmgmt.exe.
In order to prevent this issue, periodically save all Event Viewer logs. After the logs are saved, clear all the events from these logs:
If the Cisco CallManager server is integrated with the Cisco Unified Operations Manager, the high CPU issue is documented in Cisco bug ID CSCsh75491.
When Cisco Unified Operations Manager is rebooted or the daemon manager on Cisco Unified Operations Manager is restarted, with net stop crmdmgtd, net start crmdmgtd, there is a temporary spike in CPU on all monitored Windows servers. These servers include Cisco CallManager and Cisco Unity servers. The CPU spike duration depends upon the number of events recorded in these Windows Event Viewer logs:
In order to prevent this issue, periodically save all Event Viewer logs. After the logs are saved, clear all the events from these three logs.