feat: Horizontal Pod Autoscaler

[dvanmali]

Add an optional chart configuration for Horizontal Pod Autoscaler for the Surreal Deployment.

Changes chart from 0.3.6 > 0.3.7

[mumoshu]

He @dvanmali! We appreciate your contribution!

This looks great, but I believe we need to address a thing correctly to move forward.

That is, this doesn't work, until we omit the deployment's desired replicas when HPA is enabled when HPA is active. Please see

helm-charts/charts/surrealdb/templates/deployment.yaml

Line 8 in 991915b

replicas: {{ .Values.replicaCount }}

Note that it's not straightforward when you deploy the chart without HPA and then enabled it in a second deployment- It's not handled by Helm well so you need some manual steps to make it work. Please also see helm/helm#12650 for more details.

It would be great if you could:

• Fix the chart so that the deployment replicas is omitted when HPA is active
• Have some guidance in the chart README about how to migrate from non-HPA deployment to HPA-enabled

Thank you in advance for your support!

[dvanmali]

Good catch @mumoshu and thanks for reviewing this PR :)

Fixed the HPA file and added docs to state that HPA is an optional feature which can be added/removed at anytime using helm update.

[dvanmali]

@mumoshu Fixed with an enabled flag now.

I also ran helm-docs locally on my computer and some of the docs did not match, so made minor changes to values.yaml and README.md to match.

[mumoshu]

I appreciate your effort, @dvanmali!
This LGTM!

I manually verified that it supports turning HPA on/off. I've also added a quick E2E test to prevent it from breaking in the future. I hope you like it.

Having two or more pods works only when you have a distributed backend like TiKV and FoundationDB. We should document this in a separate pull request.

Thank you for your contribution, @dvanmali!

[dvanmali]

Much thanks for writing the e2e tests and the merge 🥇 Looks great! 😄