K
Q

Why GCLB's default HealthCheck interval is 5 seconds, but NEG's case is 15s?

9/19/2019

Google Cloud Load Balancing docs explains the default HealthCheck interval is 5 seconds.

https://cloud.google.com/load-balancing/docs/health-checks#create_a_health_check

also I found that NEG's default interval is 15 seconds.

https://github.com/kubernetes/ingress-gce/blob/b1d2ea77b79aa08a9de5d80881987ebf79a48876/pkg/healthchecks/healthchecks.go#L47

Why is reason? Thanks.

-- Nittaman
gke-networking
google-kubernetes-engine

Similar Questions

How to fix "Http11NioProtocol: Error reading request, ignored"
rabbitmq-peer-discovery-k8s fails. I still need to manually stop, join_cluster and start every node
Kubernetes get pods Crash
Do you need to define ports.containerPort in K8s if Dockerfile describes as EXPOSE

1 Answer

9/19/2019

The difference is due to the different native environments. Google Cloud Platform has a default health check interval of 5s. While Kubernetes has a default pod liveness check of 15s. The NEG default which you have observed is also 15s to match the pod liveness check, both coming from Kubernetes in this case.

-- TheRovinRogue
Source: StackOverflow