CI/CD with GitHub Actions

Services connected to a GitHub repository already deploy on every push. Use GitHub Actions when you want to deploy only after your tests pass, deploy several services in order, or deploy images you build yourself.

All examples use a team API token stored as the repository secret FJORDRUN_TOKEN (Settings → Secrets and variables → Actions in GitHub).

Find the IDs

Link your directory with the CLI and print the IDs:

fjordrun link
fjordrun status --json

Or read FJORDRUN_SERVICE_ID from the service's variables. Store the service ID (and the environment ID, if not production) as repository variables.

Redeploy after tests, with the CLI

# .github/workflows/deploy.yml
name: Deploy
on:
  push:
    branches: [main]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with: { node-version: 22, cache: npm }
      - run: npm ci
      - run: npm test

  deploy:
    needs: test
    runs-on: ubuntu-latest
    env:
      FJORDRUN_TOKEN: ${{ secrets.FJORDRUN_TOKEN }}
    steps:
      - name: Install the Fjordrun CLI
        run: curl -fsSL https://fjordrun.com/install.sh | sh
      - name: Deploy
        run: fjordrun deploy --project shop --service web --environment production --wait

--wait streams the build and fails the job if the deployment fails.

To avoid deploying twice (once on push, once from CI), set the service's watch paths to something that never changes, or connect the service to a branch that only CI updates.

Redeploy after tests, with curl

      - name: Deploy
        run: |
          curl -fsS -X POST \
            "https://api.fjordrun.com/api/v1/services/${{ vars.FJORDRUN_SERVICE_ID }}/deploy" \
            -H "Authorization: Bearer ${{ secrets.FJORDRUN_TOKEN }}"

Add ?environment_id=<id> for an environment other than production. The response is the new deployment; follow it with GET /api/v1/deployments/{id} until its status is success or failed.

Build the image in CI, deploy it on Fjordrun

For full control of the build (or builds that need more memory), build and push the image yourself and deploy it to a service whose source is a Docker image:

  deploy:
    needs: test
    runs-on: ubuntu-latest
    permissions: { contents: read, packages: write }
    steps:
      - uses: actions/checkout@v4
      - uses: docker/login-action@v3
        with:
          registry: ghcr.io
          username: ${{ github.actor }}
          password: ${{ secrets.GITHUB_TOKEN }}
      - uses: docker/build-push-action@v6
        with:
          push: true
          tags: ghcr.io/${{ github.repository }}:${{ github.sha }}
      - name: Point the service at the new image
        run: |
          curl -fsS -X PATCH \
            "https://api.fjordrun.com/api/v1/services/${{ vars.FJORDRUN_SERVICE_ID }}/config" \
            -H "Authorization: Bearer ${{ secrets.FJORDRUN_TOKEN }}" \
            -H "Content-Type: application/json" \
            -d '{"source": {"type": "image", "image": "ghcr.io/${{ github.repository }}:${{ github.sha }}"}}'

Changing the image of a running service deploys it. The image must be pullable by Fjordrun; make the GHCR package public, or use a public registry.

arm64

During the beta, Fjordrun's servers are arm64 (Ampere) machines. Images you build yourself must include linux/arm64: add platforms: linux/amd64,linux/arm64 to docker/build-push-action, with docker/setup-qemu-action and docker/setup-buildx-action before it. Images built by Fjordrun from your repository are built for the right architecture automatically.