Page 1 of 1

Custom vm disk utilization monitor

Posted: Sep 12, 2014 2:25 pm
by keithkleiman
I need to create a custom monitor to monitor the total disk utilization rate (ie:VMGUEST.Collect.totalDiskUsageRate) on a vm for a specific problem. Can this be done through the native SCOM monitor wizard? Can you provide any guidance or assistance with this?

Re: Custom vm disk utilization monitor

Posted: Sep 18, 2014 11:21 am
by sergey.g
Hi Keith,

That's fairly easy. Yes, you can use the built-in "Unit Monitor" wizard:
  • Select "WMI Performance Counter" as a type and "Consecutive Samples over Threshold" or "Average Threshold" sub-type (other types won't work well because disk usage is a highly volatile metric)
  • Call the monitor something like "Veeam VMware: Virtual Machine Disk Usage Analysis" (so that it can be visible in the corresponding Veeam views)
  • Specify "VMware Virtual Machine" class as a target, select Performance as a parent monitor.
  • Specify "root\nworks" as a namespace and "select totalDiskUsageRate from vmstats where entityID like '$Target/Property[Type="VeeamVirtExtensionsVMwareLibrary7001862!Veeam.Virt.Extensions.VMware.VMGUEST"]/id$''" as a query, as a frequency we do not recommend to set values less than your collection interval, so if you have 5 minutes interval - you can set frequency to 300 seconds.
  • Then you need to map performance data to a metric, specify something like VMguest-disk as an object, totalDiskUsageRate as a counter name and $Data/Property[@Name='totalDiskUsageRate']$ as a value.
  • Specify a threshold for monitor to fire and a number of consecutive samples (2 for instance, so that monitor will fire only if disk usage breached the threshold AND stayed high for 2 samples (10 minutes))
  • Configure Health settings (whether you want a monitor to go to critical or warning)
  • Then configure Alerting if you need it, you can keep the default message, but make sure you changed parameter for metric value, if you are using consecutive samples monitor - you should use $Data/Context/SampleValue$ instead of $Data/Context/Value$
And now you have it - a monitor which will fire when disk usage for a VM goes above a threshold. I've create an example of such monitor for you, you can use it as is or adjust to your own needs. ... sp=sharing

Feel free to ask any other questions about extending SCOM using data collected by Veeam MP.