What metrics does VC & DRS look at when making decisions to migrate VMs in regards to CPU utilization? Does it only look at overall CPU utilization on the ESX hosts in the cluster or does it look at other factors such as the CPU % Ready values on the hosts? For example, say you have a 16 core host and several 4 vCPU VMs that are performing poorly and have very high CPU % Ready times but the overall CPU utilization on the host is low, will DRS try to rebalance these VMs? Thanks.
I made this very request in the requests area. Right now its only looking at CPU utilization and doesn't know about the other factors with regard to CPU utilization.
I had some older hosts that while the CPU was only used at about 30% average, %RDY values were constantly 5-15 so it was becoming unacceptable and the only time DRS moved anything was when one or more of the VM's started going at 100% on my old dual socket one cores. New hardware solved that completely though (more cores) for me so its not an issue anymore for me.
I made this very request in the requests area. Right now its only looking at CPU utilization and doesn't know about the other factors with regard to CPU utilization.
I had some older hosts that while the CPU was only used at about 30% average, %RDY values were constantly 5-15 so it was becoming unacceptable and the only time DRS moved anything was when one or more of the VM's started going at 100% on my old dual socket one cores. New hardware solved that completely though (more cores) for me so its not an issue anymore for me.