A liveness probe answers one question, is this container running. A readiness probe answers a different one, can this pod serve traffic right now. Mixing them is how a slow database turns into cascading pod restarts. This one walks through the exact failure on a six-pod food delivery backend and the two-endpoint fix.
The kubelet hits /health every few seconds. A database call inside that handler is the trap, and moving dependency checks to a /ready endpoint stops the restart loop.
Do you run a database check inside your liveness probe right now? Tell us in the comments.
#Kubernetes #ReadinessProbe #LivenessProbe #K8s #DevOps #Kubelet #SRE #CloudNative #Containers #HealthCheck #KodeKloud