Jacob Straszynski

I'm thinking about this feature myself because it makes me wonder how to get accurate percentiles if we were to scale out telegraf horizontally. E.g. we could run multiple telegraf...

Cool. I'm curious what the use case here is? How are you using subsets?

It looks like it's failing a test. The strings.ToLower isn't spurious - it's there for case-insensitive lookups. I think what might be missing is a strings.ToLower in the Set method.

To follow up @jszwedko's comment, indeed this un/under-documentation still appears to be the case in April 2024. It _looks_ like the functionality is making it's way into mainline prom in...

* Is running multiple instances of kube-descheduler, each with a dedicated policy.yaml viable workaround in the short term? Any perceived gotchas with this approach? More requests on the API server...

Has this been succeeded by https://github.com/kubernetes-sigs/descheduler/issues/926? It appears the [original design doc](https://docs.google.com/document/d/1S1JCh-0F-QCJvBBG-kbmXiHAJFF8doArhDIAKbOj93I/edit#) went stale in April 2021 and was replaced by a [new document](https://docs.google.com/document/d/1tWpXoZ6YF3ksnmueM2POwASdxoiLW8y7TXuWFxT8NSo/edit#heading=h.ec817eouurjs)? If so, might I suggest closing...

I believe! Pull request would be appreciated.

Just made a comment about this on /r/php. I agree and I'm considering it.

Tests shows failures in go < 1.10. The updated client relies on `strings.Builder` which was [added to the standard library in 1.10](https://github.com/golang/go/issues/18990).

Walked a possible implementation out here: https://github.com/kubernetes/kube-state-metrics/pull/2632 Just targeted `kube_pod_status_phase` for the time being, for demonstration purposes.