Ignore prometheus alerts that are always firing in a single node setup
I took out the routes that send these alerts to the null
route:
- CPUThrottlingHigh
- KubeletTooManyPods
- KubeMemoryOvercommit
- KubeCPUOvercommit
On my cluster currently only the KubeMemoryOvercommit
and KubeCPUOvercommit
are firing so we would only need to ignore those.
It would be even better if we could exclude these alerts from prometheus so they don't even show up in prometheus and alertmanager as firing.