Skip to main content

Safe worker restart

Universal restart, zero shell execution.

Celery uses native pool_restart (graceful drain). Every other engine uses graceful self-exit via orchestrator respawn. Preflight requires both env + filesystem signals. 60-second flap guard prevents restart loops.

Preview
z4j dashboard — Safe worker restart
Restart worker web-prod-01 · celery worker engine=celery · pool=prefork · concurrency=8 PREFLIGHT ✓ Z4J_RESTART_ALLOWED=1 ✓ /tmp/z4j.allow-restart ✓ No flap in last 60s STRATEGY Native pool_restart celery-only: graceful drain STATUS ✓ Drained Respawned in 1.8s · 0 task loss [14:32:18] restart requested by alice@example.com [14:32:18] preflight OK [14:32:18] strategy: pool_restart (celery native) [14:32:18] broadcasting TERM to 8 workers... [14:32:19] 8 in-flight tasks drained [14:32:20] pool restarted, accepting new tasks [14:32:20] audit row written (id=01H6M...c42)

Mockup of the dashboard view for this feature. Live-reloading and themed to match your dashboard.

Ships with

  • Celery: native pool_restart (zero task loss)
  • Others: self-exit + orchestrator respawn (documented in-flight-loss)
  • Preflight = env signal + filesystem marker
  • 60-second flap guard
  • Never executes shell commands
Related

More capabilities

Unified action surface

One button set. Every engine.

Learn more

Stuck-task reconciliation

Tasks started forever. Fixed, automatically.

Learn more

Historical trends

Success rate, failure rate, throughput over time.

Learn more

Ready to try z4j?

Install z4j Feature docs