By Powder on Oct 21, 2011
Here are two important metrics to measure or have alerts set up on the physical host for your VM environment, to see how your VMs are performing:
· CPU ready
o Percentage of time that the virtual machine was ready, but could not get scheduled to run on the physical CPU. CPU ready time is dependent on the number of virtual machines on the host and their CPU loads.
o How long the server is waiting on a processor from the physical host machine. Longer you wait on resources, the longer it takes for your process to start. It might be easier to get a three 4-cpu machines then one 12-cpu machine. If the CPU Ready stat is above 1 then it is better to scale out than to add more CPUs to the VM.
· Disk I/O Latency
o Average amount of time, in milliseconds, to complete a SCSI command from the physical device.
o This is the time required for a computer to process a data request from the processor and then retrieve the required data from the storage device. If Latency is occurring, consider the following possibilities: changing the RAID setup, adding additional disks, cache more data into RAM, restructuring the I/O (for example, defragmentation of the disks).