Hi Gal,
some say ...
CPU Latency, i.e. %LAT_C in esxtop, includes: ready, cstp, ht busy time and effects of dynamic voltage frequency scaling and doesn't include mlmtd
https://communities.vmware.com/message/2375995
https://communities.vmware.com/thread/428764
I know you wrote that PM is disabled but can you please post Power Management Settings from your host configuration tab and also:
- output from vim-cmd hostsvc/hosthardware
- full output from ESXTOP regarding CPU resources so we can see metrics of all running VMs (for complete picture and contrast between them)
- What type apps are running on remaining VMs
I am quite convinced that your host does not have enough CPU resources to catch up with deployed VMs especially your SQL machine...
Just in addition:
Relevant counter is: average.percent.latency - Percent of time the VM is unable to run because it is contending for access to the physical CPU(s)
Other thoughts:
%LAT_C: Time VM was not scheduled due to cpu resource issue
%LAT_C: shows the CPU time unavailable to due to core sharing
Check the %LAT_C counter to see if it’s not 100%. due to two vms sharing the same core.
vCOPS
VM CPU Contention (%): Measures the difference between how much CPU the VM wants to use and how much CPU the VM gets
Thanks & Regards
P.