raft: Boot replicas proactively

This is an initial MR, which focuses on replicas that are self-bootstrapped or replicas that need to be created on a remote node.

There seems to be third case where a replica (memberID 2) is part of the cluster and later becomes the leader through Raft leader election. The routing table is updated to reflect the new leader, and the partition is closed. When a new request arrives, it is directed to replica two, which begins receiving client (Rails) requests without a partitionInfo context (which contains partitionKey and memberID). In this case, while starting the partition, we would need a way to retrieve the partitionKey and memberID, which were initially used by the replica before it became the leader. I have created an issue for that #6877

Closes #6304