feat: Possibility to Bootstrap from different node after full crash

[sebadob]

If we get into the situation where an existing cluster crashes before a new leader could have been elected, and maybe the leader volume got corrupted and cannot be recovered, a cold start after everything works again could get stuck.

Let's say we have nodes 1, 2, 3 and node 1 is the current leader.
If all nodes crash at the exact same time and the volume of node 1 get corrupted so badly, that it cannot be recovered and it happened in the middle of a log replication, so not all nodes are on the exact same log id, then nodes 2 and 3 could get into a situation where they try to re-connect to node 1 (which is dead), because they are lagging behind in log id.
If a situation like this appears, we need a way to force another node to become the new leader, basically ignoring any log id they have not received yet, even though they know that the leader had a higher one just before the crash.

This situaion is super rare, but I have been able to produce it in manual testing, even though it needed a few tries to get into it, even on purpose. However, a solution for something like this should exist.