Kubernetes Probes

Kubernetes Probes.md

• livenessProbe: tells Kubernetes whether the container is still alive. If it fails repeatedly, Kubernetes restarts the container.
• readinessProbe: tells Kubernetes whether the container is ready to receive traffic. If it fails, the pod stays running but is removed from Service load balancing.
• startupProbe: tells Kubernetes whether the app has finished starting up. While it is still failing, Kubernetes suppresses liveness/readiness handling for slow-starting apps.

Typical behavior:

• Use livenessProbe to detect deadlocks/hung processes.
• Use readinessProbe to gate traffic until dependencies/init work are done.
• Use startupProbe when startup is slow and you don’t want liveness to kill the app too early.

Simple mental model:

• liveness = “Should I restart it?”
• readiness = “Should I send traffic to it?”
• startup = “Has it booted yet?”

In practice:

• A failing livenessProbe can restart the pod.
• A failing readinessProbe will not restart it; it just marks it unready.
• A configured startupProbe gets checked first during boot; once it succeeds, normal liveness/readiness probing takes over.