I have a pretty simple question regarding how VM CPU Usage is reported on in vSphere. I've searched for the answer, but haven't found any discussions on the subject.
I don't think this specifically relates to boot storms, but that was the best way to reproduce the scenario and highlight my question. Perhaps I have a fundamental misunderstanding of what the CPU Usage metric means. And CPU Demand. Please enlighten me if so.
Graph attached.
This can mean several things including the co-stop value is rising in your host as more VMs are brought online and contend for physical CPU cores. This can also be seen if the power policy of your ESXi host is non-optimal where the host is introducing sleep states in order to throttle the CPU. Ensure your VMs are properly sized and ensure your host is using a power policy that allows ESXi to control the power state.
daphnissov,
I appreciate your reply but I still don't understand why CPU Usage is rising. The VM is not doing any more work than it had previously. I fully understand why co-stop, ready, waits, etc. would rise. But the vmware definition of CPU Usage is, "Amount of actively used virtual CPU as a percentage of total available CPU."
Why does my VM suddenly start actively using more virtual CPU at 1:14PM, when there is no additional workload on that VM? The only thing I did was start up some other VMs. I understand that the Host's CPU Usage metric would rise. But why the VM's?
