Comments
3 comments
-
Hi Idle Wild,
I would expect SQL Monitor to report similarly to what the Guest OS Perfmon is displaying - this should be the data that is being collected (see the Description of Disk avg. read/write time metrics in the Analysis graph page under Equivalent PerfMon counters) - neither of which may be correct due to how VM's affect things measured in time increments (access to CPU isn't 100% - see this article). The problem is that the counters are only running when the VM has a slice of the real CPU to run on, so that skews things that compare CPU ticks to actual clock time.
In relation to our product, I suggest you express an opinion on this with SQL Monitor's uservoice site. This open issue relates to VM monitoring.
Kind regards,
Alex -
Hi there,
Thanks for your input. I did calculate the CPU associated latency as < 1ms for the VMs in this particular datastore.
There is no particular bottleneck that I can see. Can you advise which counters Redgate uses to measure average disk latency?
Many thanks. -
Hi Idle Wild,
The disk latency metrics I can see in the Analysis graph section on the SQL Monitor website are Disk avg. read time and Disk avg. write time, which correspond to the Perfmon counters "Logical Disk: Avg. Disk sec/Read" and "Logical Disk: Avg. Disk sec/Write" respectively. Related is Disk transfers/sec (Logical Disk: Disk Transfers/sec) for total number of IOPS.
In your first post you mention the disk latency graph in SQL Monitor - what are you referring to here? Is this one of the above on the Analysis page or something from a different section?
Kind regards,
Alex
Add comment
Please sign in to leave a comment.
Redgate SQL Monitor shows the disk latency at 200+ms, while in perfmon, and elsewhere I only see disk latencies in the region of 10ms.
Can anyone advise why that might be?