Autoscaler: vpa-recommender: no way to easily debug issues with cluster state

[milesbxf]

Whilst testing my changes to improve the Prometheus history provider, it's been very difficult to verify that the right samples are being added and introspect how these samples are being linked to running pods.

I think it'd be really useful to be able to inspect the vpa-recommender's internal state. Initially a simple HTTP endpoint that returns a blob of JSON would be good, but we could potentially have some kind of UI that lets you filter and dig into things.

[kcmartin]

@milesbxf Not sure if this meets your requirements, but the VPA Recommender metrics are now exposed through kube-state-metrics:

https://github.com/kubernetes/kube-state-metrics/pull/791

[milesbxf]

@milesbxf Not sure if this meets your requirements, but the VPA Recommender metrics are now exposed through kube-state-metrics:

kubernetes/kube-state-metrics#791

Hey @kcmartin, I actually raised that PR 😁thanks though! The problem we're really having to being able to work out and understand how calculations are being made. For example - how many data points of history do we have? What are the last 10 data points?

[bskiba]

Thanks for raising, this looks useful. If anyone is willing to take this up, I'll be happy to review/discuss.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[yashbhutwala]

/remove-lifecycle stale

[yashbhutwala]

can we use the VPA checkpoint object for this and collect metrics for it in kube-state-metrics to pull info such as

• aggregation key for the resource recommendation
• first timestamp of the histogram
• total number of samples in the histogram
• percentiles (e.g. p50, p90, p99) from histograms

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[yashbhutwala]

still relevant and important.

/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.