Restore OCP instance counts in all roles after failure/termination
Use case:
- As a node operator, I need to restore minimum master/worker/infrastructure instance counts after a failure in order to maintain cluster integrity and service levels.
Definition of done:
- I can terminate any OCP instance in (e.g.) the UP node, run the playbook, and have the full set of OCP instances after the play ends.
Stretch goals:
- This works without an admin manually launching a new instance
- This works without any manual intervention by anyone, including the Platform One team (i.e., it works like an AWS auto scaling group)