Maximum CPU recommendation for a pod

[tenstad]

Basically the same as Minimum CPU recommendation for a pod, just an additional upper bound:

autoscaler/vertical-pod-autoscaler/pkg/recommender/logic/recommender.go

Line 29 in ed6ebbe

podMinCPUMillicores = flag.Float64("pod-recommendation-min-cpu-millicores", 25, `Minimum CPU recommendation for a pod`)

Which component are you using?:

vertical-pod-autoscaler

Is your feature request designed to solve a problem? If so describe the problem this feature should solve.:

VPA currently suggests cpu too close to / above CPU on node, leading to pods not starting.

Describe the solution you'd like.:

Add pod-recommendation-max-cpu-millicores and pod-recommendation-max-memory-mb flags, and cap recommendation to provided values. Maybe no upper limit by default.

Describe any alternative solutions you've considered.:

Using policies to enforce the upper bound:

autoscaler/vertical-pod-autoscaler/pkg/apis/autoscaling.k8s.io/v1/types.go

Line 205 in ed6ebbe

MaxAllowed v1.ResourceList `json:"maxAllowed,omitempty" protobuf:"bytes,4,rep,name=maxAllowed,casttype=ResourceList,castkey=ResourceName"`

Using LimitRange to enforce the upper bound: https://kubernetes.io/docs/concepts/policy/limit-range/

Additional context.:

Would be nice to define upper-bound a central place (controller), and not in policies LimitRanges.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[k8s-triage-robot]

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle rotten
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten

[tenstad]

/remove-lifecycle rotten

[adrianmoisey]

/area vertical-pod-autoscaler

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[k8s-triage-robot]

The Kubernetes project currently lacks enough active contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle rotten
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle rotten

[tenstad]

/remove-lifecycle rotten

[adrianmoisey]

This may be fixed by #7560 (which is per container, not pod)