Pod Failure Detection using K8 Liveness Probe

Kubernetes uses Liveness probe to periodically monitor the health of all pods. With this mechanism, it is easy to detect the pod failures. The K8 configuration parameter values for Liveness probe are adjusted to enable faster detection of pod failures and timely restart of faulty pods.

The configuration changes are applicable to the following pods:

  • bgpspeaker-pod

  • cache-pod

  • diameter-ep

  • dns-proxy

  • edr-monitor

  • georeplication-pod

  • gtpc-ep

  • gtpc-ep-s11

  • gtpp-ep

  • li-ep

  • nodemgr

  • oam-pod (infra-oam)

  • protocol

  • radius-ep

  • rest-ep

  • sgw-service

  • smf-service

  • udp-proxy

The early detection of pod failures by K8 enables backup or standby pod to take over message processing and achieve seamless communication.