How do readiness, liveness, and startup probes serve different purposes in Kubernetes?

Readiness indicates whether a container should receive traffic through the relevant service routing. Liveness determines whether a running container should be restarted after repeated failures. A startup probe allows a slow-starting container to initialize before liveness and readiness probing begins. Configure thresholds according to observed startup and recovery behavior. A liveness probe that fails whenever a remote database is unavailable can cause a restart storm without repairing the database. Keep probes lightweight and make readiness reflect whether the instance can safely handle the traffic it will receive.