Skip to content

docs(community): add contributor onboarding and issue templates - #49

Draft
Timur Tukaev (tym83) wants to merge 1 commit into
mainfrom
docs/contributor-onboarding
Draft

docs(community): add contributor onboarding and issue templates#49
Timur Tukaev (tym83) wants to merge 1 commit into
mainfrom
docs/contributor-onboarding

Conversation

@tym83

Copy link
Copy Markdown

Summary

Nothing in the project told a newcomer what to do first. The material exists, but
it is split in two places and neither is addressed to a first-time contributor:
the website Developer Guide
explains platform internals and per-package make targets, while
docs/agents/contributing.md
holds the actual commit, sign-off, regeneration and pull-request conventions
behind a title that reads as if it were not meant for humans.

What

contributors/README.md — a standalone walkthrough assembled from both
sources, in the order a contributor needs it: pick a task, work out which
environment the change requires, set up the fork, regenerate what is generated,
verify, commit, rebase, open the PR, and what to expect afterwards.

Three things it says that were not written down anywhere:

  • Environment cost is tiered. Charts, values, ApplicationDefinitions,
    dashboards and docs need nothing but Go and Docker. Only end-to-end work needs
    the three-QEMU-VM setup at 8 vCPU and 24 GiB each. Contributors were previously
    left to discover the 72 GiB requirement themselves, or assume it applied to
    everything.
  • Build jobs failing on a fork PR are the project's problem, not the
    contributor's.
    Forks cannot authenticate to the image registry, so any build
    job that pushes an image fails regardless of the change. Two currently approved
    external PRs are red for exactly this reason.
  • A bot comment is not a review. Every recent first-time PR was met first by
    CodeRabbit or Gemini, and several got no human response at all. The document
    states that a human answer is due within two business days and that chasing it
    is legitimate.

.github/ISSUE_TEMPLATE/governance.md and community.md — the README
routes governance, process and community matters to this repo, but the repo
offered only a blank issue. Both templates are short: what the issue is, where it
shows up, what should happen. The governance template asks for both the written
rule and the actual behaviour when the two disagree.

config.yml — two new contact links, for the onboarding guide and the
community meeting.

Why

Deliberately not a rewrite of the existing documentation. The downstream trigger
map and the gh api graphql recipe for unresolved review threads stay in
docs/agents/contributing.md and are linked, not copied — one copy, one owner.
What is duplicated here is only the material a newcomer cannot proceed without.

@gemini-code-assist

Copy link
Copy Markdown

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

@coderabbitai

coderabbitai Bot commented Aug 3, 2026

Copy link
Copy Markdown

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 5eb6e061-0f67-4f16-8296-f8e8475bf336

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

Nothing told a newcomer what to do first. The material exists but is split in two
places and neither is addressed to a first-time contributor: the website Developer
Guide explains platform internals and per-package make targets, while
docs/agents/contributing.md holds the actual commit, sign-off, regeneration and PR
conventions behind a title that reads as if it were not meant for humans.

contributors/README.md is a standalone walkthrough built from both: pick a task,
work out which environment the change actually needs (most first contributions
need no cluster at all — the three-QEMU-VM setup is only for end-to-end work), set
up the fork, regenerate what is generated, verify, commit, rebase, open the PR,
and what to expect afterwards. It states plainly that build jobs failing on fork
PRs are our registry problem rather than the contributor's, and that a bot comment
is not a review. The downstream trigger map and the review-thread query stay where
they are and are linked rather than copied.

Add issue templates for governance and community topics — the README routes both
here, but the repo offered only a blank issue.

Signed-off-by: Timur Tukaev <timur.tukaev@aenix.io>
@tym83
Timur Tukaev (tym83) force-pushed the docs/contributor-onboarding branch from dd163d3 to 6443590 Compare August 3, 2026 11:57
@IvanHunters

Copy link
Copy Markdown
Contributor

Reviewed against the actual tree (Makefile, workflows, docs/agents/contributing.md), not just the diff. Almost all of the CI/generation claims check out verbatim: the make targets, per-package show/generate, the -rd/cozyrds ApplicationDefinition layout, the area/uncategorized labeler fallback, the 3×(8 vCPU / 24 GiB) e2e VMs with the IPv4-forwarding check, both contributing.md anchors, the Assisted-By: trailer, and the fork/registry mechanism. Two things to fix before this onboards anyone:

Blocking, an undocumented commitment. "A human response is due within two business days" is not written down anywhere in the repo. ## Review Expectations (contributing.md:264) sets no response-time SLA, and the only business-day commitments are security-side with different numbers (SECURITY.md:52 3 days, docs/security/incident-response.md:34 7 days, CONTRIBUTOR_LADDER.md:154 5 days for offboarding). As written this promises a maintainer SLA the project never agreed to. Either soften it ("expect a human reply; pinging if it has been a while is legitimate") or add the rule to GOVERNANCE first and link it.

Fix, CI mechanism named wrong. "CI runs make generate across every package and then git diff --exit-code": there is no git diff --exit-code in CI. Drift is caught by codegen-drift.yml:57 (root make generate + git status --porcelain) plus the per-package run-make-generate pre-commit hook. The net effect is right, the mechanism is not, worth wording accurately since newcomers will trust it.

Non-blocking:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants