Document overprovisioning for node autoscaling

[sftim]

This is a Feature Request

What would you like to be added

• Add a new task page about configuring node overprovisioning, based on https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md#how-can-i-configure-overprovisioning-with-cluster-autoscaler

• Update https://kubernetes.io/docs/concepts/cluster-administration/cluster-autoscaling/ to link to that new page

Why is this needed

The existing document is not very easy to discover.

Comments /language en /sig autoscaling

[sftim]

The advice in https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md#how-can-i-configure-overprovisioning-with-cluster-autoscaler is actually generic to any node autoscaling solution, so no need to mention Cluster Autoscaler specifically. You do need some automatic node provisioning already configured to be able to use the technique though.

[sftim]

Also see https://github.com/kubernetes-sigs/karpenter/issues/749#issuecomment-2051561692

[sftim]

Slightly relevant to https://github.com/kubernetes/website/pull/45802

[sftim]

/triage accepted /lifecycle frozen

[afzal442]

/assign

@sftim 😄