Storage
Let's recap what we need to produce in order to monitor storage capacity:
- A line chart showing the maximum and average storage latency experienced by any VM in the datastore cluster
- A line chart showing the datastore disk capacity left in the datastore cluster
The first line chart tracks performance, while the second one tracks utilization.
Storage – performance
Just like compute, we just need to add the average super metric. The actual formula is as follows:
avg(${adaptertype=VMWARE, objecttype=VirtualMachine, attribute=virtualDisk|totalLatency, depth=2})Storage – utilization
Storage disk capacity should be tied with your actual, physical capacity. If you are using thin provisioning at the storage layer, then you need to measure it at this level. I prefer to use thin on VMware and thick on physical arrays, as management is easier from the VMware administrator's viewpoint.
Just like compute, we are using the supply-demand approach, where:
- Supply = The total datastore...
