Live migration of VM's incase of any potential failures of physical node
We have been observing virtual machines are rebooting frequently whenever there is issue with underlying physical node. This is not a good idea/option to reboot the node as the nodes are actively part of production traffic. This will leads to the inconveniences to the business and may affect.
Why don't Microsoft come up with live migration of virtual machines whenever it detects potential failure like VMWare HA does. The process of power off the VM and moving to healthy node, then powering on, not a good solution.