Hi — another small one I found while reading the await_condition path.

During leadership transfer, the node enters await_condition state. If a client command or query arrives in this window (e.g. forwarded by a follower), it hits the catch-all and gets dropped silently — the client never gets a reply back.

This adds a clause to redirect those requests to the current leader so the client can retry there instead of hanging.

BTW, that's the final issue I found. Thanks a lot!