k8s - container probes
From the docs:
Any code greater than or equal to 200 and less than 400 indicates success. Any other code indicates failure.
The liveness, readiness and startup probe are configured per container (not per Pod).
livenessProbe- why?: to know when to restart a container.
- on success: do nothing.
- on failure: restarts the container.
- default behavior: same as success
readinessProbe- why?: to know when a container is ready to accept traffic.
- on success: tells to load balancer the Pod is ready to receive requests.
- on failure: tells to load balancer to not send requests to this Pod.
- default behavior: same as success
startupProbe- why?: to know when a container app has started, and then kubelet can start running the liveness and readiness probes.
- on success: enables liveness and readiness probes.
- on failure: restarts the container
- default behavior: same as success
| probe | liveness | readiness | startup |
|---|---|---|---|
| why? | when to restart a container | when a container is ready to receive traffic | when kubelet can start running liveness/readiness probes |
| when? | after startupProbe and then continuously | after startupProbe and then continuously | after container starts / stops after success or failure |
| success | do nothing | tells the load balancer it's ready to receive traffic | tells kubelet to start running liveness/readiness probes |
| failure | container killed and restarted (see: restartPolicy) | tells load balancer to stop sending traffic | container killed and restarted (see: restartPolicy) |
| default | same as success | same as success | same as success |
diagram with the flow of all probes
