README.md

gitops repo for codeai-k8s

To change something in ArgoCD for codeai-k8s, edit this repo and merge to main.

Time for ArgoCD to notice changes to either repo:

• avg 2m 30s, max 5 minutes

Time for ArgoCD to deploy depends on what it has to restart to make the change.

TODO: time codeai deployments, including restart

Directory Structure

k8s-gitops/
  apps/
    app-of-apps/
      applicationset.yaml         # points argocd at apps/*/application.yaml and applicationset.yaml

    $app_name/
      application.yaml            # argocd will automatically find this application.yaml
      repos.yaml                  # configure application.yaml to load $app_name/*

    codeai/
      applicationset.yaml         # define argocd apps for codeai deployments: deployments/*/deployment.yaml

      deployments/
        levelbuilder/             # codeai deployment levelbuilder
          deployment.yaml         # envType=levelbuilder, targetRevision=<code-dot-org ref>
          values.yaml             # values.yaml for this deployment: dashboard_workers=27, RAILS_ENV=levelbuilder, etc
        ...

      envTypes/
        levelbuilder.values.yaml  # base values.yaml for all envType=levelbuilder
        ...

    warehouses/
      codeai/
        builds/                   # thin build-lock Freight records for Kargo
        legacy-gitflow/           # legacy branch merge metadata used as promotion gates

    kargo/
      application.yaml            # argocd app for kargo itself
      values.yaml                 # helm values for kargo install

    kargo-project-codeai/
      application.yaml            # argocd app for kargo project codeai
      project.yaml                # kargo project for codeai
      project-config.yaml         # kargo projectconfig for codeai
      warehouse.yaml              # kargo warehouse for codeai
      stages/
        levelbuilder.yaml         # kargo stage for codeai deployment levelbuilder
        ...

Bootstrap Cluster

kubectl apply -f apps/app-of-apps/applicationset.yaml