Managing complex workflows

This guide shows you how to use the advanced features of GitHub Actions, with secret management, dependent jobs, caching, build matrices, and labels.

In this article

Note: GitHub Actions support on GitHub Enterprise Server 2.22 is a limited public beta. To review the external storage requirements and request access to the beta, see "Enabling GitHub Actions and configuring storage."

Note: GitHub-hosted runners are not currently supported on GitHub Enterprise Server. You can see more information about planned future support on the GitHub public roadmap.

Overview

This article describes some of the advanced features of GitHub Actions that help you work create more complex workflows.

Storing secrets

If your workflows use sensitive data, such as passwords or certificates, you can save these in GitHub as secrets and then use them in your workflows as environment variables. This means that you will be able to create and share workflows without having to embed sensitive values directly in the YAML workflow.

This example action demonstrates how to reference an existing secret as an environment variable, and send it as a parameter to an example command.

jobs:
  example-job:
    runs-on: ubuntu-latest
    steps:
      - name: Retrieve secret
        env:
          super_secret: ${{ secrets.SUPERSECRET }}
        run: |
          example-command "$super_secret"

For more information, see "Creating and storing encrypted secrets."

Creating dependent jobs

By default, the jobs in your workflow all run in parallel at the same time. So if you have a job that must only run after another job has completed, you can use the needs keyword to create this dependency. If one of the jobs fails, all dependent jobs are skipped; however, if you need the jobs to continue, you can define this using the if conditional statement.

In this example, the setup, build, and test jobs run in series, with build and test being dependent on the successful completion of the job that precedes them:

jobs:
  setup:
    runs-on: ubuntu-latest
    steps:
      - run: ./setup_server.sh
  build:
    needs: setup
    runs-on: ubuntu-latest
    steps:
      - run: ./build_server.sh
  test:
    needs: build
    runs-on: ubuntu-latest
    steps:
      - run: ./test_server.sh

For more information, see jobs.<job_id>.needs.

Using a build matrix

You can use a build matrix if you want your workflow to run tests across multiple combinations of operating systems, platforms, and languages. The build matrix is created using the strategy keyword, which receives the build options as an array. For example, this build matrix will run the job multiple times, using different versions of Node.js:

jobs:
  build:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        node: [6, 8, 10]
    steps:
      - uses: actions/setup-node@v1
        with:
          node-version: ${{ matrix.node }}

For more information, see jobs.<job_id>.strategy.matrix.

Caching dependencies

GitHub-hosted runners are started as fresh environments for each job, so if your jobs regularly reuse dependencies, you can consider caching these files to help improve performance. Once the cache is created, it is available to all workflows in the same repository.

This example demonstrates how to cache the ~/.npm directory:

jobs:
  example-job:
    steps:
      - name: Cache node modules
        uses: actions/cache@v2
        env:
          cache-name: cache-node-modules
        with:
          path: ~/.npm
          key: ${{ runner.os }}-build-${{ env.cache-name }}-${{ hashFiles('**/package-lock.json') }}
          restore-keys: |
            ${{ runner.os }}-build-${{ env.cache-name }}-

For more information, see "Caching dependencies to speed up workflows."

Using databases and service containers

If your job requires a database or cache service, you can use the services keyword to create an ephemeral container to host the service; the resulting container is then available to all steps in that job and is removed when the job has completed. This example demonstrates how a job can use services to create a postgres container, and then use node to connect to the service.

jobs:
  container-job:
    runs-on: ubuntu-latest
    container: node:10.18-jessie
    services:
      postgres:
        image: postgres
    steps:
      - name: Check out repository code
        uses: actions/checkout@v2
      - name: Install dependencies
        run: npm ci
      - name: Connect to PostgreSQL
        run: node client.js
        env:
          POSTGRES_HOST: postgres
          POSTGRES_PORT: 5432

For more information, see "Using databases and service containers."

Using labels to route workflows

This feature helps you assign jobs to a specific self-hosted runner. If you want to be sure that a particular type of runner will process your job, you can use labels to control where jobs are executed. You can assign labels to a self-hosted runner, and then refer to these labels in your YAML workflow, ensuring that the job is routed in a predictable way.

This example shows how a workflow can use labels to specify the required runner:

jobs:
  example-job:
    runs-on: [self-hosted, linux, x64, gpu]

For more information, see "Using labels with self-hosted runners."

Next steps

To continue learning about GitHub Actions, see "Sharing workflows with your organization."

Did this doc help you?

Privacy policy

Help us make these docs great!

All GitHub docs are open source. See something that's wrong or unclear? Submit a pull request.

Make a contribution

Or, learn how to contribute.