How is a hardware failure handled by the Google compute engine

Question

what happens to my virtual instance in a scenario where hardware failure of the google compute-engine occurs?

Does the compute-engine auto migrate like it does when maintenance is performed?

Answer 1

Google states that just in case of hardware failure, your VM will be migrated if doable, or restarted on a brand new host if migration isn't doable.

Live migration keeps your instances running during:

Failed hardware like memory, CPU, network interface cards, disks, power, and so on. 

This is done on a best-effort basis; if a hardware fails totally or otherwise prevents live migration, the VM crashes and restarts mechanically and a hostError is logged.