Probes in Kubernetes: liveness, readiness, startup. What are they for and how do they differ?

In Kubernetes, there are three types of probes that help monitor the state of containers:

• Liveness probe — checks if the container is alive. If the check fails, Kubernetes restarts the container. This is useful for recovery after hangs or errors within the application.

• Readiness probe — checks if the container is ready to accept traffic. If the check fails, the container is temporarily excluded from services so that no traffic is sent to it until it becomes ready.

• Startup probe — used to verify the successful start of a container. If the application takes a long time to start, the startup probe gives more time before liveness and readiness checks begin, to avoid prematurely considering the container as failed.

Thus, these probes ensure reliability and proper load distribution in the cluster.