When should the option to leave VMs running when a host is isolated be set?

Leaving VMs running when a host is isolated is particularly important in scenarios where both virtual and physical infrastructures are designed to support high availability. This setting ensures that in case of a network or host failure, the virtual machines remain operational and accessible, either in their current state or can be quickly reassigned to other available resources within a resilient infrastructure.

When high availability is in place, the virtual machines can continue to operate despite isolation from the management network or other hosts. This minimizes downtime, allows for better resource utilization, and provides a seamless experience for users and services relying on those VMs.

In contrast, using only virtual or physical networking may create limitations in terms of access and resource management, and an isolated host might mean that critical services are rendered inaccessible. In scenarios where there is no load on the VMs, or where only one form of networking is utilized, the potential benefits of keeping the VMs running diminish, as there may not be any critical workload that needs to remain live during such events.