Investigate KubeAPIErrorBudgetBurn alerts on stackspin.net
We keep getting KubeAPIErrorBudgetBurn
on stackspin.net. Lets try to investigate the cause.
This is the apiserver_request:availability30d for stackspin.net. Even if I don't understand it fully, this graph doesn't look good, especially the writes
Unfortunatly, the kube-prometheus-stack node-exporter dashboards don't show CPU/load metrics anymore. However I installed the Node Exporter Full dashboard from the grafana dashobard hub and it shows at least CPU and load graphs, which don't show any spike that would explain the API errors starting Nov 18:
Unfortunatly, it also doesn't show disk IO.
Edited by Varac