What is the difference between a liveness probe and a readiness probe in Kubernetes?
Liveness vs readiness: a failing liveness probe triggers a container restart (the process is stuck/deadlocked). A failing readiness probe removes the pod from the service endpoint without restarting it (it is temporarily unable to serve, e.g., warming up or draining).
2 / 5
What is a startup probe in Kubernetes?
Startup probe: some applications take a long time to initialize. A startup probe prevents liveness from killing a container that is still legitimately starting. Once startup succeeds, liveness and readiness probes take over.
3 / 5
What should a deep health check verify beyond the process being alive?
Deep health check: a shallow check confirms the HTTP server responds. A deep check verifies the application can actually work — database connection is valid, cache is reachable, required secrets are loaded. This catches "running but broken" states.
4 / 5
What is the /health endpoint convention in REST services?
/health endpoint: by convention, services expose a health endpoint (often /health, /ready, /live) returning JSON with status and dependency checks. Load balancers and orchestrators poll it to determine whether to route traffic.
5 / 5
Why can an overly aggressive health check cause problems?
Aggressive checks: a health check that runs an expensive DB query every second adds load; misconfigured retry thresholds can trigger unnecessary restarts of healthy containers during a brief DB slowdown, causing a restart storm that worsens availability.