do you mean should the %RDY on vCOPS match what you see directly on the host? Yes, save for maybe some timing issues, depending on where the vCOPs is in relation to the host (i.e. one in L.A. and the other in NYC).
What version are you using?
If an issue persists, then stick with the local readings as that what the VMs will be dealing with when contention issues hit. see also:
I'm asking if what is classified as "CPU contention %" in VCOPS (5.8.4) is just another label for RDY% or something else.
Just trying to figure what is low, high or normal for this figure. I don't see actual figures for RDY or COSTOP in VCOPS.
Maybe I need to look deeper.
CPU Contention % isn't 'JUST' CPU RDY%. It includes it, but not exclusively. There is a programmatic formula to it, but essentially it considers a number of things related to contention. I consider 5-10% worth looking at for VMs just like RDY% would be. If you play around with it, you'll notice that it does line up roughly to RDY%, however you need to keep in mind that contention% is intended to be a relative statement of diff stats categories across the board, such as cpu and mem. You should be able to look at contention%=10 and think "gee that's bad" rather than needing to know exactly what perf counter and what the scenario is. Sure, specifics are relevant later, but you shouldn't need a PHD in perf troubleshooting with knowledge of every stat to see if contention exists on a VM... and that's the idea behind "contention".
# Take the vSphere Troubleshooting & Optimize and Scale courses.. they rock. I think they should be a recommended pre-req for using vC Ops at an advanced level since it gets you in the right frame of mind.
Thanks, that's basically what I was looking for, what percentage is considered problematic.
The odd thing is that I see a number of vms with high contention percentages but this isn't reflected in the hosts they are running on.
The percentage for the hosts that the vms are on are 7 and 9 %.