Heartbeats are used to ensure that the host is up and running. If no heartbeats are received within a configured timeframe, the host is considered down, and the HA event will occur, restarting the VMs running on the unresponsive host.

There are two kind of heartbeat:

• Network heartbeat: The master node will periodically check if the slave node is available over the network using ICMP pings.
• Storage heartbeat: Each ESXi hypervisor stores an empty file on the shared datastore, which is exclusively locked by the ESXi hypervisor. If the file is not locked, it means that the hosts have lost access to the storage.