Dashboards for the storage team
Chapter 11, SDDC Key Counters, explains how distributed storage, such as VSAN, differs fundamentally from centralized storage. Chapter 14, Storage Counters, shows that this results in different objects and counters. This chapter shows an example of how we monitor them.
LUN performance monitoring
A common requirement of storage monitoring is to see the data at the physical storage level. The data at this level eliminates the information at VM level. If the performance is good at this level but poor at the VM level, you should check the disk queue length inside the Guest OS.
ESXi tracks the latency on each device it sees; this means LUNs (in central storage) or physical disks (in distributed storage). For central storage, this means VMFS datastores. There are no devices in NFS datastores.
For distributed storage, ESXi has to be able to see the physical disk. If the disk is directly passed through to the VM, it is no longer visible by the vmkernel. The vmkernel...
