Steve Biedermann 2305660d5a fix: exit shards when simulation threads fail ... Supervise the shard simulation thread and exit with status 1 when it panics or stops, allowing the container restart policy to recover instead of leaving a live control server with no simulation. Cover startup panics and actor-reply timeouts with subprocess regression tests, and document recovery behavior. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>		2026-10-21 20:00:05 +02:00
..
cluster.rs	fix: exit shards when simulation threads fail	2026-10-21 20:00:05 +02:00