Sometimes, some guest VM's IO latency become high(more than 1300). At this time, IOPS is very low.
I have refferd KB (ARTICLE NUMBER 000001636: NCC Health Check: vm_io_latency_check) from support portal, but I cannot image in my environment as just some guest VM's IO latency is high.
What do you think that there is any condition or situation?
Best answer by rreiter
We experienced this in our environment and after a lot of investigation by Nutanix and VMware, it was determined that gflag changes had to be made to alter the timing and efficiency of the oplog draining. This was done by Nutanix support with the assistance of their engineers.