Deploy Apache Airflow DAGs to Astro

Deploy to Astro

This GitHub action automates deploying code from your GitHub repository to a Deployment on Astro, Astronomer's data orchestration platform and managed service for Apache Airflow.

You can use and configure this GitHub action to easily deploy Apache Airflow DAGs to an Airflow environment on Astro. Specifically, you can:

Avoid manually running astro deploy with the Astro CLI every time you make a change to your Astro project.
Automate deploying code to Astro when you merge changes to a certain branch in your repository.
Incorporate unit tests for your DAGs as part of the deploy process.
Create/delete a Deployment Preview. A Deployment Preview is an Astro Deployment that mirrors the configuration of your original Deployment.

This GitHub action runs as a step within a GitHub workflow file. When your CI/CD pipeline is triggered, this action:

Checks out your GitHub repository.
Optionally creates or deletes a Deployment Preview to test your code changes on before deploying to production.
Checks whether your commit only changed DAG code.
Optional. Tests DAG code with pytest. See Run tests with pytest.
Either runs:
- astro deploy --dags if the commit only includes DAG code changes,
- or astro deploy (as well as astro dev parse) if the commit includes any non-DAG-code-related changes.

Prerequisites

To use this GitHub action, you need:

An Astro project. See Create a project.
A Deployment on Astro. See Create a Deployment.
An Organization, Workspace, or Deployment API Token. See API Tokens
Or (deprecated) a Deployment API key ID and secret. See Deployment API keys.

Tip

Astronomer recommends using GitHub Actions secrets to store ASTRO_API_TOKEN or Deployment API Keys. See the example in Workflow file examples.

Use this action

To use this action, read Automate code deploys with CI/CD. You will:

Create a GitHub Actions workflow in your repository that uses the latest version of this action. For example, astronomer/[email protected].
Configure the workflow to fit your team's use case. This could include creating a deployment preview or adding tests. See Configuration options.
Make changes to your Astro project files in GitHub and let this GitHub Actions workflow take care of deploying your code to Astro.

Tip

Astronomer recommends setting up multiple environments on Astro. See the Multiple branch GitHub Actions workflow in Astronomer documentation.

Configuration options

The following table lists the configuration options for the Deploy to Astro action.

Name	Default	Description
`action`	`deploy`	Specify what action you would like to take. Use this option to create or delete deployment previews. Specify either `deploy`, `create-deployment-preview`, `delete-deployment-preview` or `deploy-deployment-preview`. If using `deploy` or `deploy-deployment-preview` one should also specify `deploy-type`.
`deploy-type`	`infer`	Specify the type of deploy you would like to do. Use this option to deploy images and/or DAGs or DBT project. Possible options are `infer`, `dags-only`, `image-and-dags` or `dbt`. `infer` option would infer between DAG only deploy and image and DAG deploy based on updated files. For description on each deploy type checkout deploy type details
`deployment-id`	`false`	Specifies the id of the deployment you to make a preview from or are deploying too.
`deployment-name`	`false`	Specifies The name of the deployment you want to make preview from or are deploying too. Cannot be used with `deployment-id`
`description`		Configure a description for a deploy to Astro. Description will be visible in the Deploy History tab.
`root-folder`	`.`	Path to the Astro project, or dbt project for dbt deploys.
`parse`	`false`	When set to `true`, DAGs are parsed for errors before deploying to Astro. Note that when an image deploy is performed (i.e. `astro deploy`), parsing is also executed by default. Parsing is not performed automatically for DAG-only deploys (i.e. `astro deploy --dags`).
`pytest`	`false`	When set to `true`, all pytests in the `tests` directory of your Astro project are run before deploying to Astro. See Run tests with pytest
`pytest-file`	(all tests run)	Specifies a custom pytest file to run with the pytest command. For example, you could specify `/tests/test-tags.py`.
`force`	`false`	When set to `true`, your code is deployed and skips any pytest or parsing errors.
`image-name`		Specifies a custom, locally built image to deploy. To be used with `deploy-type` set to `image-and-dags` or `infer`
`workspace`		Workspace id to select. Only required when `ASTRO_API_TOKEN` is given an organization token.
`preview-name`	`false`	Specifies custom preview name. By default this is branch name “_” deployment name.
`checkout`	`true`	Whether to checkout the repo as the first step. Set this to false if you want to modify repo contents before invoking the action. Your custom checkout step needs to have `fetch-depth` of `0` and `ref` equal to `${{ github.event.after }}` so all the commits in the PR are checked out. Look at the checkout step that runs within this action for reference.
`deploy-image`	`false`	If true image and DAGs will deploy for any action that deploys code. NOTE: This option is deprecated and will be removed in a future release. Use `deploy-type: image-and-dags` instead.
`build-secrets`	``	Mimics docker build --secret flag. See https://docs.docker.com/build/building/secrets/ for more information. Example input 'id=mysecret,src=secrets.txt'.
`mount-path`	``	Path to mount dbt project in Airflow, for reference by DAGs. Default /usr/local/airflow/dbt/{dbt project name}
`checkout-submodules`	`false`	Whether to checkout submodules when cloning the repository: `false` to disable (default), `true` to checkout submodules or `recursive` to recursively checkout submodules. Works only when `checkout` is set to `true`. Works only when `checkout` is set to `true`.
`wake-on-deploy`	`false`	If true, the deployment will be woken up from hibernation before deploying. NOTE: This option overrides the deployment's hibernation override spec.

Outputs

The following table lists the outputs for the Deploy to Astro action.

Name	Description
`preview-id`	The ID of the created deployment preview. Only works when action=create-deployment-preview.

Deploy Type Details

The following section describe each of the deploy type input value in detail to avoid any confusion:

infer: In this mode, deploy-action would run through all the file changes:

if there are no file changes in the configured root-folder then it skips deploy
if there are changes only in dags/ folder, then it will do a dags deploy
otherwise it would do a complete image deploy

image-and-dags: In this mode, deploy-action would run through all the file changes:

if there are no file changes in the configured root-folder then it skips deploy
otherwise it would do a complete image deploy

dags-only: In this mode, deploy-action would run through all the file changes:

if there are no file changes in the configured root-folder then it skips deploy
if all the file changes are in folders except dags folder then it skips deploy
otherwise it would do a dag only deploy

dbt: In this mode, deploy-action would run through all the file changes:

if there are no file changes in the configured root-folder then it skips deploy
otherwise it would do a dbt deploy

Workflow file examples

In the following example, the GitHub action deploys code to Astro. This example assumes that you have one Astro Deployment and one branch. When a change is merged to the main branch, your Astro project is deployed to Astro. DAG files are parsed on every deploy and no pytests are ran.

name: Astronomer CI - Deploy code

on:
  push:
    branches:
      - main

env:
  ## Set API Token as an environment variable
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Deploy to Astro
      uses: astronomer/[email protected]
      with:
        deployment-id: <deployment id>
        parse: true

Use the following topics to further configure the action based on your needs.

Change the Root folder

In the following example, the folder /example-dags/ is specified as the root folder.

steps:
- name: Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    root-folder: /example-dags/

Run Pytests

In the following example, the pytest located at /tests/test-tags.py runs before deploying to Astro.

steps:
- name: Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    pytest: true
    pytest-file: /tests/test-tags.py

Ignore parsing and testing

In the following example, force is enabled and both the DAG parse and pytest processes are skipped.

steps:
- name: Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    force: true

Deploy a custom Docker image

In the following example, a custom Docker image is built and deployed to an Astro Deployment.

name: Astronomer CI - Additional build-time args

on:
  push:
    branches:
      - main

jobs:
  build:
    runs-on: ubuntu-latest
    env:
      ## Set API Token as an environment variable
      ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}
    steps:
    - name: Check out the repo
      uses: actions/checkout@v3
    - name: Create image tag
      id: image_tag
      run: echo ::set-output name=image_tag::astro-$(date +%Y%m%d%H%M%S)
    - name: Build image
      uses: docker/build-push-action@v2
      with:
        tags: ${{ steps.image_tag.outputs.image_tag }}
        load: true
        # Define your custom image's build arguments, contexts, and connections here using
        # the available GitHub Action settings:
        # https://github.com/docker/build-push-action#customizing .
        # This example uses `build-args` , but your use case might require configuring
        # different values.
        build-args: |
          <your-build-arguments>
    - name: Deploy to Astro
      uses: astronomer/[email protected]
      with:
        deployment-id: <deployment id>
        deploy-type: image-and-dags
        image-name: ${{ steps.image_tag.outputs.image_tag }}

Deploy a DBT project

In the following example we would be deploying the dbt project located at dbt folder in the Github repo

steps:
- name: DBT Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    deploy-type: dbt
    root-folder: dbt

Deploy DAGs and DBT project from same repo

In the following example we would setup a workflow to deploy dags/images located at astro-project and dbt deploy from dbt project located at dbt folder in the same Github repo

steps:
- name: DBT Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    deploy-type: dbt
    root-folder: dbt
- name: DAGs/Image Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    root-folder: astro-project/
    parse: true

Wake on deploy

In the following example, the deployment is woken up from hibernation before deploying via the wake-on-deploy option.

steps:
- name: Deploy to Astro
  uses: astronomer/[email protected]
  with:
    deployment-id: <deployment id>
    wake-on-deploy: true

This is especially useful when you have a deployment that is hibernated and you want to deploy to it. This option overrides the deployment's hibernation override spec.

Deployment Preview Templates

This section contains four workflow files that you will need in your repository to have a full Deployment Preview Cycle running for your Deployment. A Deployment Preview is an Astro Deployment that mirrors the configuration of your original Deployment. This Deployment Preview can be used to test your new pipelines changes before pushing them to your original Deployment. The scripts below will take your pipeline changes through the following flow:

When a new branch is created a Deployment Preview will be created based off your original Deployment
When a PR is created from a branch code changes will be deployed to the Deployment Preview
When a PR is merged into your "main" branch code changes will be deployed to the original Deployment
When a branch is deleted the corresponding Deployment Preview will also be deleted

Create Deployment Preview

name: Astronomer CI - Create deployment preview

on:
  create:
    branches:
    - "**"

env:
  ## Sets Deployment API key credentials as environment variables
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Create Deployment Preview
      uses: astronomer/[email protected]
      with:
        action: create-deployment-preview
        deployment-id: <orginal deployment id>

Deploy to Deployment Preview

name: Astronomer CI - Deploy code to Preview

on:
  pull_request:
    branches:
      - main

env:
  ## Sets Deployment API key credentials as environment variables
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Deploy to Deployment Preview
      uses: astronomer/[email protected]
      with:
        action: deploy-deployment-preview
        deployment-id: <orginal deployment id>

DBT Deploy to Deployment Preview

name: Astronomer - DBT Deploy code to Preview

on:
  pull_request:
    branches:
      - main

env:
  ## Sets Deployment API key credentials as environment variables
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Deploy to Deployment Preview
      uses: astronomer/[email protected]
      with:
        action: deploy-deployment-preview
        deploy-type: dbt
        deployment-id: <orginal deployment id>
        root-folder: dbt

Delete Deployment Preview

name: Astronomer CI - Delete Deployment Preview

on:
  delete:
    branches:
    - "**"

env:
  ## Sets Deployment API key credentials as environment variables
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Delete Deployment Preview
      uses: astronomer/[email protected]
      with:
        action: delete-deployment-preview
        deployment-id: <orginal deployment id>

Deploy to Original Deployment

name: Astronomer CI - Deploy code to Astro

on:
  push:
    branches:
      - main

env:
  ## Sets Deployment API key credentials as environment variables
  ASTRO_API_TOKEN: ${{ secrets.ASTRO_API_TOKEN }}

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
    - name: Deploy to Astro
      uses: astronomer/[email protected]
      with:
        deployment-id: <orginal deployment id>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

GitHub Action

Deploy Apache Airflow DAGs to Astro