search

cloud-gov / kubernetes-broker

Broker for kubernetes based services
Apache License 2.0
5 stars 6 forks source link

Redis health check #47

Closed jmcarp closed 7 years ago

jmcarp commented 7 years ago

Shamelessly lifted from https://github.com/kubernetes/charts/blame/master/stable/redis/templates/deployment.yaml.

Will verify and revise once I'm home and back on the work vpn.

jmcarp commented 7 years ago

I actually read TFD and confirmed that failing the liveness probe will cause a container restart: https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-probes/. Failing the readiness probe will take the pod out of the k8s load balancer, which should be useful once we get around to adding HA.

cnelson commented 7 years ago

What's the failure mode when there are no nodes in the k8s load balancer pool?

jmcarp commented 7 years ago

Looks like connections to a service with no pods times out. Since this story is just about restarting an unhealthy redis instance, I think I'll just drop the readiness probe for now.