Damien Grisonnet comments

Results 522 comments of


                                            Damien Grisonnet

self metrics leak Go pointer types to the output

That is reasonable to me for the following reasons: - from an ingestion standpoint, it is backward compatible to update label values - these are self metrics, so ksm repo...

self metrics leak Go pointer types to the output

> This is not a breaking change. A chart like If we take this literally, it is a breaking change since it could create issues that were not present before...

self metrics leak Go pointer types to the output

So my tl;dr is that this particular change is very unlikely to be a breaking change, but it could always be worth mentioning it in the changelog.

Deprecate VPA

I am still very much against cutting a new major version of ksm just for deprecating VPA metrics. This is not an overhaul of the codebase, we are just deprecating...

Deprecate VPA

There is a doc explaining how to get them back in https://github.com/kubernetes/kube-state-metrics/blob/main/docs/customresourcestate-metrics.md#verticalpodautoscaler

do not create annotation and label metrics if none are allowed

> The backwards incompatible change is a concern though, and I am not sure if it's sufficient to just add a entry in the changelog. I wouldn't be too concerned...

do not create annotation and label metrics if none are allowed

@juliantaylor do you want to take a stab at implementing this new feature? I am happy to give some pointers if needed.

[WIP] Add metrics of kube_pod_status_ready_time and kube_pod_status_containers_ready_time

Since @liangyuanpeng expressed their intention to push this effort over the finish line I'd rather wait a bit more for them to get the time they need to work on...

kube_pod_container_resource.* showing duplicate metrics with and without node label

IMO we shouldn't do anything about this. The goal of ksm is to reflect the state of the kube-apiserver. In this particular case, the pod object was created but wasn't...

kube_pod_container_resource.* showing duplicate metrics with and without node label

> i would like to point out though that this is not what we're experiencing. in other words, we do see an overlap in time when both metrics exist. maybe...