Hi,
we are wondering if it is possible to tell from various performance counters if our cluster is overloaded (too many vms per host). Also can someone please tell me which cpu is meant (physical or virtual) when a CPU alert is triggered on a VM.
Regards,
Oliver
Hi,
You can use the built in vSphere Performance tab for monitoring the performance of your cluster, VMs, Host.
There are 3rd party monitoring tools as well such as Veeam One
If your VM is having CPU alert that means the virtual CPU is overloaded and not the Physical one, the physical CPU information you can get from the Summary tab of your ESXi host. You can also check within the VM it's CPU utilization if it's a Windows VM then just check the task manager
Hope this information is useful
Nikhil
Hi,
thanks for the repsonse. In terms of overloaded cluster, is there a recommended value of vms per host? We have 4 IBM HS22 (48GB, 2 CPU sockets, 4 cores + HT enabled) running in a DRS Cluster. Currently we have 37 VMs running.
Check out the top 25 Free tools presentation by David Davis and Kendrick Coleman.
There should be a couple of good tools to get you headed in the right direction.
http://www.youtube.com/watch?v=vWnb6BMj9U4
Damian Karlson has a good post on trouble shooting CPU and Memory performance.
http://damiankarlson.com/vcap-dca4-exam/objective-6-2-troubleshoot-cpu-and-memory-performance/
See knowledge base KB: 2002181 Converting between CPU summation and CPU % ready values
Another good ESXTOP guide by Simon Greaves.
http://simongreaves.co.uk/blog/esxtop-guide
I hope that helps.
Roger Lund
Well, I base it on CPU ready.
But, you need to make sure you have the cores to support those vms etc.
I would never recommend someone getting into virtualization or starting out to try a super high consolidation ratio without a tool like xangati, vkernel, or foglight.
Roger l
Sent from my iPad