Avoid accidental live DB upgrades if other management servers are in Up state

[borisstoyanov]

problem

A user might execute an incident upgrade when adding a new management server with later version(if let's say there's a new minor version meanwhile in the same upstream repo).

To avoid this, I propose when we identify an upgrade path is due we:

1. Check DB if there's any other management server in Up state
2. If yes, we skip the DB upgrade and shut down the management server.

versions

any, detected with 4.20.1 env and recent 4.20.2 release

The steps to reproduce the bug

1. Have an env running 4.20.1
2. Add new management server, with version 4.20.2
3. Start the management server
   ...

What to do about it?

No response

[DaanHoogland]

marked for 20.3 as it is an improvement, but if DB changes are needed, as they might be, I think we should move it to milestone 23.