I've been observing something unexpected. The VEEAM Monitor VM (dedicated W2k3R2 SP2, 2 vCPU, 3GB memory, FC storage, SQL 2008R2 Express) is generating the most IOPS of all 250+ VM's in our environment.
This is measured by looking at Datastore Total I/O average over the past day. The numbers I see are about 250 IOPS, roughly 2/3 are read, about 1/3 are write IOPS.
Interestingly the write activity is fairly consistant over the last 7 days (~100 IOPS), during which the read IOPS are fairly low (less than 10). But last 2 or 3 days the read IOPS have been much higher, over 200 for hours at a time, during which time the write IOPS drop considerably to around 25.
I've been unable to pinpoint exactly which process is generation the disk io, the only processes with cpu activity are sqlserver.exe and veeamdcs.exe
If anyone could take a look in his or her environment and reports what number your seeing as to establish a baseline.
Any help in this issue would be greatly appreciated.