Prep initial charts indexing

[scottrigby]

Fixes #7, Fixes #8

• Cherry-pick commits from the one PR that was merged since the code was copied [stable/prometheus] update prometheus to 2.20.1 and cm reloader to 0.4.0 helm/charts#23506 (see Hold all PRs in stable repo for prometheus* charts #13 for context)
• Deprecate charts/prometheus-operator (see Deprecate prometheus-operator chart before helm repo index, so that it won't be listed in the hubs #8 and Rename prometheus-operator #9 for context)
• Remove prow OWNERS files (see Invite current stable prometheus chart maintainers as collaborators with write access #12 for context)

Charts that need to be prepared for initial helm index

• charts/prometheus
• charts/prometheus-adapter
• charts/prometheus-blackbox-exporter
• charts/prometheus-cloudwatch-exporter
• charts/prometheus-consul-exporter
• charts/prometheus-couchdb-exporter
• charts/prometheus-mongodb-exporter
• charts/prometheus-mysql-exporter
• charts/prometheus-nats-exporter
• charts/prometheus-node-exporter
• charts/prometheus-postgres-exporter
• charts/prometheus-pushgateway
• charts/prometheus-rabbitmq-exporter
• charts/prometheus-redis-exporter
• charts/prometheus-snmp-exporter
• charts/prometheus-to-sd

Suggestion: per the steps in the plan at prometheus-community/community#28 (comment) (which has been approved for the prometheus-community), for this PR review and comments let's discuss only what should happen in the intial step. To be clear, this PR is the first step, before "Step 2. Collaborate with prometheus devs to refactor/rename and simplify charts as appropriate." I outlined what I think we need in #7, but please let me know if I'm missing anything.

📊 Maintainers poll

How do you all feel about the new charts/prometheus README update in b324685? I Needed to update references to stable repo, but took the opportunity to reorganize, fix and simplify. Is this a good pattern for the other chart READMEs? Note major differences are:

• Added Helm 2 and Helm 3 commands, with links to Helm 3 command reference links (from there, users can select Helm 2 from the version dropdown if they wish)
• Removed the options table that's duplicated from (often out of sync with) values.yaml in favor of helm inspect (Helm 2) and helm show values (Helm 3)

Results: new template has been approved by all who responded

[Miouge1]

LGTM, I think it makes sense to drop the table of values from the README.md, it's often hard to keep it in sync with reality.

[scottrigby]

So far it sounds like we may be able to compromise with helm upgrade [RELEASE_NAME] [CHART] --install. Welcome other opinions here as well (though not trying to bike-shed). Looking for feedback on the readme reorganization overall as a pattern for the rest of the chart updates. Thanks!

[monotek]

I'm ok with using the upgrade command. Maybe the long version (--install instead of -i) of the arguments should be used to be clear.

Rest of the readme looks good to me and could be used as a template for the other charts.

[scottrigby]

@monotek @steven-sheehy & gang, I put the Helm 3 commands above Helm 2. And instead of using helm upgrade for installation instructions, I took suggestions above for making helm install use an explicit release name, then added helm upgrade commands in a similar way to the "Upgrading chart" section so there's no confusion. I also reorganized the upgrading MAJOR version headers slightly to be more clear to end users, depending on the version they're upgrading to. This has worked out well for other community charts. The pattern is like this:

## Upgrading Chart
(helm commands)

## To v6.0
(instructions)

## To v5.0
(instructions)

## To v4.0
(instructions)

## From `foo` legacy chart located <HERE> (if any)
(instructions)

Does this look reasonable? I can apply this same pattern to the other charts.

[scottrigby]

I've now resolved all prometheus* charts maintainer feedback on the initial prometheus chart README changes, so will apply this pattern to the other charts so we can get this merged and expedite the repo transfer ⏳

[scottrigby]

PS re-running gh action job. This appears to be an API blip:

✖︎ prometheus-consul-exporter => (version: "0.1.6", path: "charts/prometheus-consul-exporter") > Error validating maintainer 'timm088': 502 Bad Gateway

Just tested locally and it validates fine:

$ docker run --rm -it -v $(pwd):/app:cached quay.io/helmpack/chart-testing:v3.0.0
$ / #
$ / # cd /app
$ /app # ct lint --charts /app/charts/prometheus-consul-exporter/
...
1 chart(s) linted, 0 chart(s) failed
------------------------------------------------------------------------------------------------------------------------
 ✔︎ prometheus-consul-exporter => (version: "0.1.7", path: "/app/charts/prometheus-consul-exporter/")
------------------------------------------------------------------------------------------------------------------------
All charts linted successfully

However, now getting API throttling errors:

✖︎ prometheus => (version: "11.12.1", path: "charts/prometheus") > Error validating maintainer 'gianrubio': 429 too many requests

According to GitHub everything is fine https://www.githubstatus.com/ 🤔

We may have to just try to run the job again 🤦 ⏳

[scottrigby]

😑 https://github.com/scottrigby/prometheus-helm-charts/runs/1007003791

I've honestly no idea which GH user GITHUB_TOKEN ties to for GH API rate limits. Don't seem to see this in the docs for this token in actions. It doesn't seem to be my user though:

X-RateLimit-Limit: 5000
X-RateLimit-Remaining: 4992
X-RateLimit-Reset: 1597920256

Fine, I'll go back to sleep and re-run later 😴

PS: If anyone wants to join us at the Deep Dive Into Helm session today, it starts in a little over 2 hours 🕣 See some of you then!

Edit: temporarily bypassed the issue by setting ct lint --validate-maintainers option to false. So far so good, let's see if anything else is broken on ct install. Once any further install test issues are fixed, we should be good to go here. Thanks for the reviews everyone! 💖

[scottrigby]

Woot! tests have passed! Let's get to the next steps. Not all maintainers of these charts have participated yet, but they may or may not be available. The goal of this PR is to change at least the charts readmes so the initial helm repo indexing will contain only updated info, and relevant to the new repo (we had to fix things along the way only because some of these charts have not been touched for some time). In future PRs, they will only be merged once at least one maintainer of the affected charts approve the PR. For that reason I recommend future PRs try to touch only one chart at a time like we do in the stable and incubator git source repo (github.com/helm/charts).

[scottrigby]

Update: Maintainers please check #11 for another poll to see if there are any remaining blockers for the repo transfer to prometheus-community

There are also testing steps in that issue, if anyone wants to try it out now 🚀

[gkarthiks]

approved