Alarms based on counters such as "Virtual machine disk usage (disk.usage.average)" may not be triggered for virtual machines (VMs) residing on vSAN datastores.
The followings are observed:
vSAN (All versions)
This behavior is a limitation due to the vSAN architecture.
In standard datastores such as VMFS, performance statistics are collected at the SCSI layer of the ESXi storage stack. However, virtual disks (vDisks) on vSAN bypass the SCSI layer and communicate via vSAN-specific protocols. Consequently, values are not reflected in conventional counters like disk.usage.average.
Performance data in vSAN environments is collected and managed separately by the vSAN Performance Service, rather than the standard vSphere performance statistics.
To monitor performance or configure alarms for VMs on vSAN datastores, use one of the following alternative methods: