Closed marquesj2-ppb closed 5 months ago
Well, that's not good. To confirm, are you on the latest version?
Also, what metric are you querying ip-10-159-137-250.eu-west-1.compute.internal
in the first picture?
There seems to be an issue around this part of the code. I will try to replicate it with Kind.
https://github.com/jmcgrath207/k8s-ephemeral-storage-metrics/blob/master/main.go#L321
Yes, installed yesterday in deployment mode.
ephemeral_storage_node_percentage on the first picture. Node capacity and node available also have the same behaviour, but didn't include it in the print.
Tried running it as daemonset but couldn't get the metric scrapping to work, will try to take a look closely today on that.
So I can't replicate this issue with Kind since I can't scale nodes. However, I can scale up and down in Minikube.
Once I get CI testing working for Minikube, I will investigate this issue again.
@marquesj2-ppb I have fixed this bug and added some e2e testing around it in the new release.
https://github.com/jmcgrath207/k8s-ephemeral-storage-metrics/releases/tag/1.6.0
Let me know how it goes
I'll try it out next week and let you know! Thanks a lot for the quick fix!
Can confirm, 1.6 fixes the issue
Thanks for testing @lozbrown
Closing this issue @marquesj2-ppb, but feel free to re-ping me if you still see this issue.
Hey π
Seeing a strange behaviour when deploying this application.
Node is removed, but keeps pushing metrics that get picked up by prometheus.