An administrator has a three-host cluster with vSphere HA enabled. The master host is unable to ping one of the slave hosts over the management network. The slave host in question is showing as not re

vMotion is used for live migration of running VMs and plays no role in vSphere HA's failure detection logic or VM restart decision-making.

Datastore heartbeating being enabled would assist the HA master in confirming whether the host has truly failed or is merely isolated, and would more likely facilitate VM restarts rather than prevent them.

vSphere HA's VM Monitoring feature depends entirely on VMware Tools to supply in-guest heartbeat signals that allow the HA agent to determine whether virtual machines are responsive. Without VMware Tools installed on any of the VMs, HA has no visibility into individual VM health and therefore cannot trigger automatic VM restarts, even when the host appears unresponsive on the management network.

VM Monitoring being disabled would prevent VM-level restart triggers, but in a host-level unresponsiveness scenario like this, the host failure path - not the VM Monitoring path - would normally initiate restarts, so disabled VM Monitoring alone does not explain the absence of restarts.