How argoproj/argo-helm works
Project type and purpose
• This is an open-source Helm chart repository (“Argo Helm Charts”) that packages, versions, and publishes Kubernetes manifests (CRDs and resources) for several Argo projects (argo-cd, argo-workflows, argo-events, argo-rollouts, argocd-apps, argocd-image-updater).
• The primary purpose is to maintain community-driven Helm charts, run CI checks (lint, test, publish), and host them on GitHub Pages / OCI registry for users to install into their Kubernetes clusters.
File structure
• Top-level folders:
• charts/ – each subdirectory is a Helm chart (argo-cd, argo-events, etc.).
• scripts/ – helper scripts for docs generation, linting, release analysis, version bumps.
• .github/ – GitHub workflows for CI (lint/test/publish), issue templates, dependabot.
• Root config files (renovate.json, CODEOWNERS, Security, CONTRIBUTING, README).
• Within each charts/<chart-name>/:
• Chart.yaml, values.yaml – metadata & defaults
• templates/ – Kubernetes manifest templates grouped by component (controller, server, CRDs, etc.)
• crds/ (sometimes) vs. templates/crds/ – custom resource definitions
• ci/ – chart-specific test values
• Scripts under scripts/release-analysis/ gather and analyze upstream releases.
README for additional insights
• Confirms that CRDs are managed under templates to allow upgrades.
• References to Helm Capabilities (conditional templating based on Kubernetes API versions).
• Instructions for users integrating the charts via Helm CLI, helmfile, or Kustomize.
a. Main components:
• GitHub Repository (root of charts, scripts, workflows)
• CI/CD Pipelines (GitHub Actions: lint-and-test, publish)
• Chart Modules (each chart folder with sub-components)
• OCI Registry / GitHub Pages (chart hosting endpoint)
• Users / Tooling (Helm CLI, helmfile, Kustomize)
• Kubernetes Clusters (target environment where charts are deployed)
• External Upstream Projects (argo-cd, argo-workflows, etc., for CRD sources)
b. Relationships & interactions:
• Developer → pushes commits → CI workflows run lint/test → on success generate chart packages → publish to OCI/GitHub Pages.
• User tools (helm CLI, helmfile, Kustomize) pull charts from the registry → render templates (with .Capabilities) → install into Kubernetes cluster → results in K8s resources (Deployments, Services, CRDs).
• scripts/release-analysis periodically pull upstream release data → generate analytics.
c. Architectural patterns & design principles:
• Modularization: one chart per Argo project, each chart subdivides templates by function (controllers, webhooks, server, CRDs).
• CI-as-Code: GitHub Actions define lint/test/publish stages.
• Template-Driven Deployment: leverage Helm’s templating engine and built-in Capabilities object.
• Separation of Concerns: CRDs vs. non-CRDs; chart values vs. templates vs. scripts.
d. Key technologies & libraries:
• Helm (Capabilities, chart hooks, templating)
• Kubernetes API objects (CRDs, Deployments, Services, RBAC)
• GitHub Actions for CI/CD
• Python scripts & Mega-Linter for release analysis
• Renovate for dependency management
Elements
• Clear labels: e.g., “GitHub Repo”, “Helm Chart: argo-cd”, “GitHub Action: publish”, “Helm CLI”, “K8s Cluster”
• Commits → CI workflows
• CI workflows → Chart registry
• Helm CLI → Registry → Kubernetes
• scripts → release-analysis reports
• Purple for external services (OCI registry, Kubernetes cluster)