We have around 500 plus virtual servers running on VSphere. They are all Windows Server 2008 R2 and are Citrix servers. We have seen alerts that come like every day on different servers with 100% CPU utilization at different times on different servers. Sometimes, it is the same servers. There are different process that causes the spike. So, it is not consistent with one particular process that is causing this spike. The spike can be for a short while or it can be for several minutes. We want to narrow it down to find out what could be causing this issue, whether it is a PROCESS or it is something in the OS or VMWare side?
What can we do in order to find it. Microsoft ran some perflogs and analyzed it and said that these are the processes that are causing spikes. OK, but, why are they spiking. If it different on different servers then why is different processes causing spikes. If it is one particular process that is causing, then I can say, ok this application owner / vendor needs to be contacted in order to fix this, but it is not consistent. So, how to proceed with this kind of issue? The alerts are set to 88% CPU usage in VMware.
Appreciate any input you all could provide in regard to this.
Thank in advance!
There are few things to check to get the information on CPU utilization. Run esxtop and capture the cpu ready time, cpu used % and cpu utilization percentage..Is there any limit or reservation sets on VM, Are these vm's are part of any resource configured cluster or resource pool. Is there any specific time when these value are going high?