Quentin Bisson

I was wrong, this happened in kube-state-metrics 2.10 https://github.com/kubernetes/kube-state-metrics/releases/tag/v2.10.0 (cf. the top message)

Regarding the fix I would assume this would be needed under `operations/mimir`? In that case I'm not sure I would be the best to fix it because I'm definitely lost...

@dimitarvdimitrov is there an issue to add hpa to the components that do not have it yet like the ingester?

I just find this and we would really be interested in this

@dimitarvdimitrov is there a way to help push this forward? I really think this would provide a lot of value

@duricanikolic do you know if it's possible to get this PR forward? This is a really important feature

@cstyan I would really appreciate the review :)

Sure, so the issue is that if we set the per_cluster_label now to `cluster_id` then all occurrences of `cluster=$cluster` are perfectly replaced to `cluster_id=$cluster` but when the englobing replace them...

The previous one should be enough but i'll run it an test. From what I understood, there are 4 use cases: - cluster=$cluster - cluster=~$cluster - cluster in templating or...

@cstyan I checked and there is no other use cases for cluster apart from those 3 replace. The one used in the templating section is managed via the addCluster libsonnet...