logo hsb.horse
← Back to glossary index

Glossary

Liveness

Liveness is an indicator of whether a process or service is still alive rather than stopped, hung, or stuck in an unrecoverable state.

Published: • Updated:
#ops #kubernetes #health-check #process

Translations

German French Japanese Korean Portuguese

Liveness is an indicator used to check whether a process or service is dead, hung, or stuck in a state that requires a restart.

An HTTP server may still have an open port, but if it is internally deadlocked or no longer making progress on certain processing, liveness is considered lost. While readiness asks whether the service can accept traffic, liveness asks whether it is still alive.

Typical use cases

  • Kubernetes liveness probe
  • Dead/alive checks in monitoring systems
  • Restart decisions for automated recovery

Notes

If you try to express a lost liveness state only with readiness, it becomes easy to miss failures that should trigger a restart. Treating them as separate roles is safer.