That’s an interesting process… you could improve it with some ansible - if that’s your thing… or use snapshots on the VM(s) and roll back?

That’s kinda what I’m doing with this (physical obviously) Pi… take a full backup now and again… do upgrades… rollback when completely borked.