search

kubernetes-monitoring / kubernetes-mixin

A set of Grafana dashboards and Prometheus alerts for Kubernetes.
Apache License 2.0
2.08k stars 597 forks source link

`code_verb:apiserver_request_total:increase30d` does not perform sum by cluster #916

Open lorenzofelletti opened 4 months ago

lorenzofelletti commented 4 months ago

Whilst all rules in kube-apiserver-availability.rules group performs a sum by cluster operation, code_verb:apiserver_request_total:increase30d does not.

I'm referring to this specifically:

          {
            record: 'code_verb:apiserver_request_total:increase%s' % SLODays,
            expr: |||
              avg_over_time(code_verb:apiserver_request_total:increase1h[%s]) * 24 * %d
            ||| % [SLODays, $._config.SLOs.apiserver.days],
          },

Is that a bug, or is there a reason for it?