When a host running production workloads can become inaccessible or fails due to an unplanned event, AHV will restart the VMs on other hosts automatically (similar to ESXi). For tracking purposes, it is important to get a list of VMs that were running on the failed node and restarted across other nodes in the cluster.

One way to gather this information is via the Prism web page. Each time there is an HA event, a task is generated in Prism that will give you a list of VMs that were restarted.

Another method is to use the command line interface to look at the tasks and match the output UUID with the UUIDs of the VMs in Acropolis.

More information about the above two methods can be found in KB 8956.