Federated Hpa feature doc

[irfanurrehman]

cc @kubernetes/sig-federation-pr-reviews @kubernetes/sig-federation-feature-requests
cc @quinton-hoole

---

This change is 

[k8sio-netlify-preview-bot]

Deploy preview ready!

Built with commit 30df752

https://deploy-preview-5487--kubernetes-io-vnext-staging.netlify.com

[zacharysarah]

@irfanurrehman 👋 I'd like to make some small edits. Would you please make sure that the box to allow edits from maintainers is checked?

[irfanurrehman]

@irfanurrehman I'd like to make some small edits. Would you please make sure that the box to allow edits from maintainers is checked?

@zacharysarah apologies for missing it while raising the PR. Done now.

[ghost]

Review status: 0 of 3 files reviewed at latest revision, 6 unresolved discussions.

---

docs/tasks/administer-federation/hpa.md, line 2 at r1 (raw file):

---
title: Federated Hpas

I'd spell it out: "Federated Horizontal Pod Autoscalers (HPA)"

---

docs/tasks/administer-federation/hpa.md, line 6 at r1 (raw file):

{% capture overview %}
This guide explains how to use Hpas in the Federation control plane.

HPA's ?

---

docs/tasks/administer-federation/hpa.md, line 65 at r1 (raw file):

Once a federated Hpa is created, the federation control plane will partition and
create an Hpa in all underlying Kubernetes clusters.

Please mention cluster selectors to restrict the clusters into which the HPA gets federated.

---

docs/tasks/administer-federation/hpa.md, line 118 at r1 (raw file):

## Deleting a Federated Hpa

You can delete a federated Hpa as you would delete a Kubernetes

I think it's worth mentioning that this will delete all underlying HPA's, if cascading deletion is used.

---

docs/tasks/administer-federation/hpa.md, line 131 at r1 (raw file):

To a `federation user` interacting with federated control plane (or simply federation),
the interaction is same as interacting with a normal k8s cluster (but with a limited set of

'almost identical' rather than 'the same'.

---

docs/tasks/administer-federation/hpa.md, line 146 at r1 (raw file):

they are needed most, or in other words where the load is more. Federated hpa feature
achieves this by manipulating the min and max replicas on the hpas it creates in the
federated clusters. It reduces the max and min replicas from those clusters which does

I think you need to be clearer here that the federation does not actually monitor the CPU utilization itself, but rather relies on the underlying cluster HPA's to do the auto-scaling, and only re-assigns min and max replicas between clusters in an attempt to prevent individual clusters from running up against their max replica limits. A brief example might help to clarify this.

---

Comments from Reviewable

[ghost]

I made a few suggested edits. None of them are showstoppers, and can be addressed in a followup PR if necessary.

[tengqm]

then -> than

[tengqm]

greater -> greater than

[tengqm]

are greater then -> is greater than

[tengqm]

are lesser then -> is less than

[tengqm]

apis, that which are federated -> APIs that are federated

[irfanurrehman]

I have updated the doc to fix the review comments.

---

Review status: 0 of 3 files reviewed at latest revision, 11 unresolved discussions.

---

docs/tasks/administer-federation/hpa.md, line 2 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

I'd spell it out: "Federated Horizontal Pod Autoscalers (HPA)"

Done.

---

docs/tasks/administer-federation/hpa.md, line 6 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

HPA's ?

made hpa -> HPA at all places in the doc.

---

docs/tasks/administer-federation/hpa.md, line 65 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

Please mention cluster selectors to restrict the clusters into which the HPA gets federated.

Done.

---

docs/tasks/administer-federation/hpa.md, line 79 at r1 (raw file):

Previously, tengqm (Qiming) wrote…

then -> than

Done.

---

docs/tasks/administer-federation/hpa.md, line 80 at r1 (raw file):

Previously, tengqm (Qiming) wrote…

greater -> greater than

Done.

---

docs/tasks/administer-federation/hpa.md, line 88 at r1 (raw file):

Previously, tengqm (Qiming) wrote…

are greater then -> is greater than

Updated the sentence, please check if still sounds wrong.

---

docs/tasks/administer-federation/hpa.md, line 90 at r1 (raw file):

Previously, tengqm (Qiming) wrote…

are lesser then -> is less than

Updated the sentence, please check if still sounds wrong.

---

docs/tasks/administer-federation/hpa.md, line 118 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

I think it's worth mentioning that this will delete all underlying HPA's, if cascading deletion is used.

Done.

---

docs/tasks/administer-federation/hpa.md, line 131 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

'almost identical' rather than 'the same'.

Done.

---

docs/tasks/administer-federation/hpa.md, line 132 at r1 (raw file):

Previously, tengqm (Qiming) wrote…

apis, that which are federated -> APIs that are federated

Done.

---

docs/tasks/administer-federation/hpa.md, line 146 at r1 (raw file):

Previously, quinton-hoole (Quinton Hoole) wrote…

I think you need to be clearer here that the federation does not actually monitor the CPU utilization itself, but rather relies on the underlying cluster HPA's to do the auto-scaling, and only re-assigns min and max replicas between clusters in an attempt to prevent individual clusters from running up against their max replica limits. A brief example might help to clarify this.

Done.

---

Comments from Reviewable

[irfanurrehman]

@quinton-hoole @zacharysarah
I think this PR should wait for kubernetes/kubernetes#49950 to get merged.
I believe that would happen when the code freeze is lifted.

[ghost]

The PR kubernetes/kubernetes#49950 has merged. All edits /lgtm. This can be merged.

[steveperry-53]

I'll start a docs review now.

[steveperry-53]

@irfanurrehman See 30df752 for suggested edits.

[steveperry-53]

@irfanurrehman Ping. As far as I know, we're still shooting for a 9/27 release.

[ghost]

Thanks @steveperry-53 . Yes, still aiming at 9/27 release. Perhaps @irfanurrehman was unaware that he was responsible for applying your edits. Is the documentation process that we should follow documented somewhere? Any pointers much appreciated.

[zacharysarah]

@quinton-hoole 👋

Perhaps @irfanurrehman was unaware that he was responsible for applying your edits.

Generally, docs contributors provide reviews and perform small edits.* In the case of more extensive edits, docs contributors provide feedback for a PR owner to implement. In the case of this PR specifically, I edited 30df752 last week; Steve's feedback referenced that PR.

Is the documentation process that we should follow documented somewhere?

Not in terms of PR review. Looking at the README and the contribution guidelines, there's definitely room to set clearer expectations. I've opened an issue to improve the review process: #5620

For now, we need either:

• @irfanurrehman to push a commit that incorporates updates
• @irfanurrehman to un-check and re-check the box allowing edits from maintainers

[*]The GitHub UI has a known issue: edits from maintainers are allowed only if the PR owner un-checks and re-checks the "allow edits from maintainers" box after the PR is created.

[steveperry-53]

@quinton-hoole @zacharysarah @irfanurrehman, We don't need anyone to check the Allow edits box, and we don't need @irfanurrehman to push anything. I already pushed a commit with suggested edits. All we need is for someone to agree to what I've done in 30df752.

[ghost]

@steveperry-53 Yes, thanks, 30df752 looks fine.

/lgtm

Let me know if you need us to do anything else.

[steveperry-53]

Reviewed 2 of 3 files at r1, 1 of 1 files at r4.
Review status: all files reviewed at latest revision, 11 unresolved discussions.

---

Comments from Reviewable

[steveperry-53]

@quinton-hoole Thanks for the lgtm.

@quinton-hoole, @zacharysarah Can one of you help get the code-review/reviewable to be green. To this day, I am mystified by Reviewable. Thanks. And if this isn't easy, I'm happy to merge this without a green check on the Reviewable line.

[steveperry-53]

---

Review status: all files reviewed at latest revision, 12 unresolved discussions.

---

Comments from Reviewable