"Implementation of Virtual Machines in Availability Sets with Managed Disks would have provided resiliency against significant service impact for VM based workloads. ". So if you'd implemented your service properly then that's exactly what would have happened.