Running variations of jobs in a workflow

Create a matrix to define variations for each job.

In this article

About matrix strategies

A matrix strategy lets you use variables in a single job definition to automatically create multiple job runs that are based on the combinations of the variables. For example, you can use a matrix strategy to test your code in multiple versions of a language or on multiple operating systems.

Using a matrix strategy

Use jobs.<job_id>.strategy.matrix to define a matrix of different job configurations. Within your matrix, define one or more variables followed by an array of values. For example, the following matrix has a variable called version with the value [10, 12, 14] and a variable called os with the value [ubuntu-latest, windows-latest]:

jobs:
  example_matrix:
    strategy:
      matrix:
        version: [10, 12, 14]
        os: [ubuntu-latest, windows-latest]

A job will run for each possible combination of the variables. In this example, the workflow will run six jobs, one for each combination of the os and version variables.

By default, GitHub will maximize the number of jobs run in parallel depending on runner availability. The order of the variables in the matrix determines the order in which the jobs are created. The first variable you define will be the first job that is created in your workflow run. For example, the above matrix will create the jobs in the following order:

  • {version: 10, os: ubuntu-latest}
  • {version: 10, os: windows-latest}
  • {version: 12, os: ubuntu-latest}
  • {version: 12, os: windows-latest}
  • {version: 14, os: ubuntu-latest}
  • {version: 14, os: windows-latest}

A matrix will generate a maximum of 256 jobs per workflow run. This limit applies to both GitHub-hosted and self-hosted runners.

The variables that you define become properties in the matrix context, and you can reference the property in other areas of your workflow file. In this example, you can use matrix.version and matrix.os to access the current value of version and os that the job is using. For more information, see "Accessing contextual information about workflow runs."

Example: Using a single-dimension matrix

You can specify a single variable to create a single-dimension matrix.

For example, the following workflow defines the variable version with the values [10, 12, 14]. The workflow will run three jobs, one for each value in the variable. Each job will access the version value through the matrix.version context and pass the value as node-version to the actions/setup-node action.

jobs:
  example_matrix:
    strategy:
      matrix:
        version: [10, 12, 14]
    steps:
      - uses: actions/setup-node@v4
        with:
          node-version: ${{ matrix.version }}

Example: Using a multi-dimension matrix

You can specify multiple variables to create a multi-dimensional matrix. A job will run for each possible combination of the variables.

For example, the following workflow specifies two variables:

  • Two operating systems specified in the os variable
  • Three Node.js versions specified in the version variable

The workflow will run six jobs, one for each combination of the os and version variables. Each job will set the runs-on value to the current os value and will pass the current version value to the actions/setup-node action.

jobs:
  example_matrix:
    strategy:
      matrix:
        os: [ubuntu-22.04, ubuntu-20.04]
        version: [10, 12, 14]
    runs-on: ${{ matrix.os }}
    steps:
      - uses: actions/setup-node@v4
        with:
          node-version: ${{ matrix.version }}

A variable configuration in a matrix can be an array of objects.

matrix:
  os:
    - ubuntu-latest
    - macos-latest
  node:
    - version: 14
    - version: 20
      env: NODE_OPTIONS=--openssl-legacy-provider

This matrix produces 4 jobs with corresponding contexts.

- matrix.os: ubuntu-latest
  matrix.node.version: 14
- matrix.os: ubuntu-latest
  matrix.node.version: 20
  matrix.node.env: NODE_OPTIONS=--openssl-legacy-provider
- matrix.os: macos-latest
  matrix.node.version: 14
- matrix.os: macos-latest
  matrix.node.version: 20
  matrix.node.env: NODE_OPTIONS=--openssl-legacy-provider

Example: Using contexts to create matrices

You can use contexts to create matrices. For more information about contexts, see "Accessing contextual information about workflow runs."

For example, the following workflow triggers on the repository_dispatch event and uses information from the event payload to build the matrix. When a repository dispatch event is created with a payload like the one below, the matrix version variable will have a value of [12, 14, 16]. For more information about the repository_dispatch trigger, see "Events that trigger workflows."

{
  "event_type": "test",
  "client_payload": {
    "versions": [12, 14, 16]
  }
}

on:
  repository_dispatch:
    types:
      - test
 
jobs:
  example_matrix:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        version: ${{ github.event.client_payload.versions }}
    steps:
      - uses: actions/setup-node@v4
        with:
          node-version: ${{ matrix.version }}

Expanding or adding matrix configurations

Use jobs.<job_id>.strategy.matrix.include to expand existing matrix configurations or to add new configurations. The value of include is a list of objects.

For each object in the include list, the key:value pairs in the object will be added to each of the matrix combinations if none of the key:value pairs overwrite any of the original matrix values. If the object cannot be added to any of the matrix combinations, a new matrix combination will be created instead. Note that the original matrix values will not be overwritten, but added matrix values can be overwritten.

For example, this matrix:

strategy:
  matrix:
    fruit: [apple, pear]
    animal: [cat, dog]
    include:
      - color: green
      - color: pink
        animal: cat
      - fruit: apple
        shape: circle
      - fruit: banana
      - fruit: banana
        animal: cat

will result in six jobs with the following matrix combinations:

  • {fruit: apple, animal: cat, color: pink, shape: circle}
  • {fruit: apple, animal: dog, color: green, shape: circle}
  • {fruit: pear, animal: cat, color: pink}
  • {fruit: pear, animal: dog, color: green}
  • {fruit: banana}
  • {fruit: banana, animal: cat}

following this logic:

  • {color: green} is added to all of the original matrix combinations because it can be added without overwriting any part of the original combinations.
  • {color: pink, animal: cat} adds color:pink only to the original matrix combinations that include animal: cat. This overwrites the color: green that was added by the previous include entry.
  • {fruit: apple, shape: circle} adds shape: circle only to the original matrix combinations that include fruit: apple.
  • {fruit: banana} cannot be added to any original matrix combination without overwriting a value, so it is added as an additional matrix combination.
  • {fruit: banana, animal: cat} cannot be added to any original matrix combination without overwriting a value, so it is added as an additional matrix combination. It does not add to the {fruit: banana} matrix combination because that combination was not one of the original matrix combinations.

Example: Expanding configurations

For example, the following workflow will run four jobs, one for each combination of os and node. When the job for the os value of windows-latest and node value of 16 runs, an additional variable called npm with the value of 6 will be included in the job.

jobs:
  example_matrix:
    strategy:
      matrix:
        os: [windows-latest, ubuntu-latest]
        node: [14, 16]
        include:
          - os: windows-latest
            node: 16
            npm: 6
    runs-on: ${{ matrix.os }}
    steps:
      - uses: actions/setup-node@v4
        with:
          node-version: ${{ matrix.node }}
      - if: ${{ matrix.npm }}
        run: npm install -g npm@${{ matrix.npm }}
      - run: npm --version

Example: Adding configurations

For example, this matrix will run 10 jobs, one for each combination of os and version in the matrix, plus a job for the os value of windows-latest and version value of 17.

jobs:
  example_matrix:
    strategy:
      matrix:
        os: [macos-latest, windows-latest, ubuntu-latest]
        version: [12, 14, 16]
        include:
          - os: windows-latest
            version: 17

If you don't specify any matrix variables, all configurations under include will run. For example, the following workflow would run two jobs, one for each include entry. This lets you take advantage of the matrix strategy without having a fully populated matrix.

jobs:
  includes_only:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        include:
          - site: "production"
            datacenter: "site-a"
          - site: "staging"
            datacenter: "site-b"

Excluding matrix configurations

To remove specific configurations defined in the matrix, use jobs.<job_id>.strategy.matrix.exclude. An excluded configuration only has to be a partial match for it to be excluded. For example, the following workflow will run nine jobs: one job for each of the 12 configurations, minus the one excluded job that matches {os: macos-latest, version: 12, environment: production}, and the two excluded jobs that match {os: windows-latest, version: 16}.

strategy:
  matrix:
    os: [macos-latest, windows-latest]
    version: [12, 14, 16]
    environment: [staging, production]
    exclude:
      - os: macos-latest
        version: 12
        environment: production
      - os: windows-latest
        version: 16
runs-on: ${{ matrix.os }}

Note: All include combinations are processed after exclude. This allows you to use include to add back combinations that were previously excluded.

Example: Using an output to define two matrices

You can use the output from one job to define matrices for multiple jobs.

For example, the following workflow demonstrates how to define a matrix of values in one job, use that matrix in a second jobs to produce artifacts, and then consume those artifacts in a third job. Each artifact is associated with a value from the matrix.

YAML
name: shared matrix
on:
  push:
  workflow_dispatch:

jobs:
  define-matrix:
    runs-on: ubuntu-latest

    outputs:
      colors: ${{ steps.colors.outputs.colors }}

    steps:
      - name: Define Colors
        id: colors
        run: |
          echo 'colors=["red", "green", "blue"]' >> "$GITHUB_OUTPUT"

  produce-artifacts:
    runs-on: ubuntu-latest
    needs: define-matrix
    strategy:
      matrix:
        color: ${{ fromJSON(needs.define-matrix.outputs.colors) }}

    steps:
      - name: Define Color
        env:
          color: ${{ matrix.color }}
        run: |
          echo "$color" > color
      - name: Produce Artifact
        uses: actions/upload-artifact@v4
        with:
          name: ${{ matrix.color }}
          path: color

  consume-artifacts:
    runs-on: ubuntu-latest
    needs:
    - define-matrix
    - produce-artifacts
    strategy:
      matrix:
        color: ${{ fromJSON(needs.define-matrix.outputs.colors) }}

    steps:
    - name: Retrieve Artifact
      uses: actions/download-artifact@v4
      with:
        name: ${{ matrix.color }}

    - name: Report Color
      run: |
        cat color

Handling failures

You can control how job failures are handled with jobs.<job_id>.strategy.fail-fast and jobs.<job_id>.continue-on-error.

jobs.<job_id>.strategy.fail-fast applies to the entire matrix. If jobs.<job_id>.strategy.fail-fast is set to true or its expression evaluates to true, GitHub will cancel all in-progress and queued jobs in the matrix if any job in the matrix fails. This property defaults to true.

jobs.<job_id>.continue-on-error applies to a single job. If jobs.<job_id>.continue-on-error is true, other jobs in the matrix will continue running even if the job with jobs.<job_id>.continue-on-error: true fails.

You can use jobs.<job_id>.strategy.fail-fast and jobs.<job_id>.continue-on-error together. For example, the following workflow will start four jobs. For each job, continue-on-error is determined by the value of matrix.experimental. If any of the jobs with continue-on-error: false fail, all jobs that are in progress or queued will be cancelled. If the job with continue-on-error: true fails, the other jobs will not be affected.

jobs:
  test:
    runs-on: ubuntu-latest
    continue-on-error: ${{ matrix.experimental }}
    strategy:
      fail-fast: true
      matrix:
        version: [6, 7, 8]
        experimental: [false]
        include:
          - version: 9
            experimental: true

Defining the maximum number of concurrent jobs

By default, GitHub will maximize the number of jobs run in parallel depending on runner availability. To set the maximum number of jobs that can run simultaneously when using a matrix job strategy, use jobs.<job_id>.strategy.max-parallel.

For example, the following workflow will run a maximum of two jobs at a time, even if there are runners available to run all six jobs at once.

jobs:
  example_matrix:
    strategy:
      max-parallel: 2
      matrix:
        version: [10, 12, 14]
        os: [ubuntu-latest, windows-latest]