Resilience: add restart policy (unless-stopped) so the server self-heals after a kill

The container died with exit 137 (external SIGKILL, OOMKilled=false, host had 428GiB
free) — not an OOM. With RestartPolicy=no a single kill left it permanently down.
--restart unless-stopped auto-recovers from crashes/kills/daemon restarts while still
respecting a deliberate docker stop.