When the machine hosting customer VMs is down for maintenance — for upgrading the OS or the hardware, what happens to the customer VMs? Are they “live migrated” to another host, or are they stopped (or paused) during the maintenance? If they are “live migrated”, does it include the migration of data on attached volumes?
For the past few years, I’ve been using platforms supporting live migration of workloads and network-attached block storage, which seems maintenance operations have been pretty transparent, except perhaps for a small slowdown during migration, and a very short “freeze” period during the switchover from the old to the new machine.