CPU95PercentOnAverage fires for nodes that are no longer part of cluster

[OlleLarsson]

Describe the bug Read title

To Reproduce

1. Remove node from cluster
2. wait
3. observe alert firing

Expected behavior Nodes that have been removed should be excluded from the query

Screenshots

Version (add all relevant versions):

• v0.25.0

Additional context Looks like the nodes that have been removed still shows up in this query due to the long time range and their values will be very small.

rate(node_cpu_seconds_total{job="node-exporter",mode="idle"}[1d])

[Xartos]

Is this only an issue for a day or will the alert fire for a longer period?

[OlleLarsson]

It will be for one day (i assume)