Push Helm charts from CI pipelines to IAC repos and monitor Kubernetes deployments in real time.

DAG is a GitOps deployment automation system. It receives Helm charts from CI pipelines, commits them to your Infrastructure-as-Code repository, and monitors the Kubernetes cluster until the deployment succeeds or fails — streaming status updates back to your pipeline in real time via SSE.

┌─────────────────┐         HTTP          ┌─────────────────┐
│   CI Pipeline   │ ──────────────────▶   │   DAG Server    │
│  (dag-deploy)   │ ◀────────────────── │                 │
│                 │         SSE           │                 │
└─────────────────┘                       └────────┬────────┘
                                                   │
                                    ┌──────────────┼──────────────┐
                                    ▼              ▼              ▼
                              ┌──────────┐  ┌──────────┐  ┌──────────┐
                              │ IAC Repo │  │  GitLab/ │  │   K8s    │
                              │  (Git)   │  │  GitHub  │  │ Cluster  │
                              └──────────┘  └──────────┘  └──────────┘

• GitOps Native — commits Helm charts to your IAC Git repo, keeping Git as the single source of truth. DAG never applies changes to Kubernetes directly.
• Real-Time Monitoring — streams deployment status via SSE so your CI pipeline shows live progress from pending through deployed.
• Flux & Plain Helm — monitors both FluxCD HelmRelease CRDs and plain Helm release secrets.
• CI Auto-Detection — automatically picks up repo URL, job ID, and job token from GitLab CI and GitHub Actions environments.
• Concurrent Safety — serializes deployments to the same IAC repo with mutex locks to prevent merge conflicts.

• Node.js 22+ (or use the Docker images)
• PostgreSQL 16+
• A Kubernetes cluster with FluxCD or plain Helm
• An IAC Git repository

Server:

docker pull ghcr.io/zyno-io/dag/server:latest

CLI:

docker pull ghcr.io/zyno-io/dag/cli:latest
# OR
npm install -g @zyno-io/dag-cli

dag-deploy ./chart \
  --server https://dag.example.com \
  --deploy-version $CI_COMMIT_SHA \
  --set image.tag=$CI_COMMIT_SHA

The CLI packages the chart, submits it to the server, and streams deployment status until completion.

Submit a Helm chart for deployment.

dag-deploy <chart-path> [options]

Accepts a chart directory, .tgz, or .tar.gz. Values overrides (--values-file, --set, --set-json, --set-file) require a chart directory. Exits 0 on success, 1 on failure.

deploy:
  stage: deploy
  image: ghcr.io/zyno-io/dag/cli:latest
  script:
    - dag-deploy ./chart
      --server https://dag.example.com
      --deploy-version $CI_COMMIT_SHA
      --set image.tag=$CI_COMMIT_SHA

Repo URL, job ID, and job token are auto-detected from GitLab CI environment variables.

deploy:
  runs-on: ubuntu-latest
  container:
    image: ghcr.io/zyno-io/dag/cli:latest
  steps:
    - uses: actions/checkout@v4
    - name: Deploy
      run: |
        dag-deploy ./chart \
          --server https://dag.example.com \
          --deploy-version ${{ github.sha }} \
          --set image.tag=${{ github.sha }}
      env:
        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

helm repo add dag https://signal24.github.io/dag
helm install dag dag/dag \
  --set postgres.host=your-pg-host \
  --set postgres.password=your-pg-password \
  --set config.cryptoSecret=your-secret

docker run -d \
  -e APP_ENV=production \
  -e PG_HOST=localhost -e PG_DATABASE=dag -e PG_USER=dag -e PG_PASSWORD_SECRET=secret \
  -p 3000:3000 \
  ghcr.io/zyno-io/dag/server:latest

See the full documentation for all options including observability settings.

Full documentation is available at zyno-io.github.io/dag.

MIT