fix: ensure worker graceful shutdown is tried in cluster pool