Operating NFs — readiness probes, graceful shutdown, hot reload

The Three Kubernetes Probes

Kubernetes uses three probe types to manage pod lifecycle. startupProbe runs first and gates liveness and readiness checks during slow startup — useful for NFs that need time to register with NRF, load configuration, or warm caches. livenessProbe tests whether the process is alive; failure triggers a restart. readinessProbe tests whether the pod can serve traffic; failure removes it from Service load-balancing without restarting. The distinction matters: an NF that has stalled (deadlock) wants a liveness restart; an NF that is alive but temporarily can't serve traffic (NRF disconnected,…