Prometheus-operator: kube-prometheus: Mismatch between dashboards and rules

[mikebryant]

https://github.com/coreos/prometheus-operator/blob/9d4f8feb906d465008a0ea45fb94782c3f90446a/contrib/kube-prometheus/assets/prometheus/rules/kubernetes.rules#L134 has cluster_resource_verb:apiserver_latency:quantile_seconds

https://github.com/coreos/prometheus-operator/blob/9d4f8feb906d465008a0ea45fb94782c3f90446a/contrib/kube-prometheus/assets/grafana/kubernetes-control-plane-status-dashboard.json#L401 references apiserver_latency_seconds:quantile

I can't find a definition for apiserver_latency_seconds:quantile, so the dashboard doesn't work.

(Similar issue for all the graphs that aren't ...UP in that dashboard)

[brancz]

Yes, you're right this happened due to some syncronization issues with internal extensions to these rules, I'll try to figure out how to solve this properly. Sorry about this, I'll try to make sure there is a smooth solution to this now and in the future.

[brancz]

We've made the decision to open source core rules for Kubernetes clusters, so I will soon be synchronizing them back into this repository so they will be consistent with alerts and dashboards.

[arslanbekov]

@brancz hello, no news on this issue?

[brancz]

Sorry about, this I should be able to get to this very soon.