Validate DAGs, Variables and Dependencies before deploying to production by creating an isolated Airflow Environment on Docker Container with supplied variables and dependencies.

Supports Airflow 2.x and Airflow 3.x.

Want to test Airflow DAGs in tests/dags with plugins in tests/plugins, requirements in tests/requirements.txt, variables in tests/var.json, and connections in tests/conns.json?

• Provide your dependency files (requirements.txt) to test your Python dependencies
• Your var.json to test your variables
• Your conns.json to test your connections
• Path to your DAGs directory to import and validate DAGs with supplied dependencies and variables
• Path to your DAG plugins directory to test DAGs using plugins
• Boolean flag for whether to load example DAGs or not

- name: 'Validate DAGs'
  uses: jayamanikharyono/airflow-dag-action@v3
  with:
    dagPaths: dags
    pluginPaths: plugins
    requirementsFile: requirements.txt
    varFile: var.json
    connFile: conns.json
    accessToken: ${{ secrets.GITHUB_TOKEN }}
    airflowVersion: "2.10.4"
    airflowExtras: "async,postgres,cncf.kubernetes"

- name: 'Validate DAGs'
  uses: jayamanikharyono/airflow-dag-action@v3
  with:
    dagPaths: dags
    pluginPaths: plugins
    requirementsFile: requirements.txt
    varFile: var.json
    connFile: conns.json
    accessToken: ${{ secrets.GITHUB_TOKEN }}
    airflowVersion: "3.1.7"
    airflowExtras: "cncf.kubernetes"
    validationRules: "all"
    enableSarif: "true"

Result

Airflow 2 Sample Comments Airflow 3 Sample Comments

• Structured Markdown PR comments with DAG tables, errors, and warnings
• PR comment deduplication — each job updates its own comment instead of creating duplicates
• File-level GitHub Actions annotations on PR diffs
• SARIF output for GitHub Code Scanning integration with line-level precision
• Diff-aware validation — only validates directories with changed files (enabled by default)
• Multiple DAG directory validation (comma-separated dagPaths)
• Custom pytest validation scripts
• Warning-only mode (failOnError: "false")
• Pip dependency caching support
• Pre-built Docker images via GHCR
• JSON results output for programmatic downstream use
• Improved error classification — import errors are automatically categorized into sub-rules (cycle, duplicates, connection, schedule, config) based on exception patterns
• Load duration tracking in validation summary
• Constraint URL validation with clear error messages for invalid Airflow versions

PR comments are rendered from a Jinja2 template and include:

• Status badge (pass/fail) with workflow and job context
• DAG summary table (DAG ID, file, task count, owner, schedule)
• Error and warning tables with rule tags
• Collapsible full error details for long tracebacks
• Collapsible custom test results
• Collapsible environment details (variables, connections, plugins)
• Footer with applied rules and directories checked

- name: 'Validate DAGs'
  id: validate
  uses: jayamanikharyono/airflow-dag-action@v3
  with:
    dagPaths: dags
    accessToken: ${{ secrets.GITHUB_TOKEN }}
    enableSarif: "true"

- name: Upload SARIF
  if: always() && steps.validate.outputs.sarif-file != ''
  uses: github/codeql-action/upload-sarif@v4
  with:
    sarif_file: validation_results.sarif
    category: airflow-dag-validation

Pre-built images are published to GHCR to speed up CI runs. The build matrix covers:

• The action no longer passes inputs as positional Docker args. All inputs are now read from environment variables (INPUT_*). This is transparent if you use the action via uses: as intended.
• The validator-result output value is now "pass" or "fail" (previously the raw output of the validation step).
• Minimum recommended Docker image is Python 3.11.

Add the new airflowVersion input to pin your Airflow version explicitly. The default is "2.10.4", so existing Airflow 2 workflows will continue to work without changes.

- uses: jayamanikharyono/airflow-dag-action@v3
  with:
    dagPaths: dags
    accessToken: ${{ secrets.GITHUB_TOKEN }}
    airflowVersion: "2.10.4"        # explicit (was implicit in v2)
    airflowExtras: "async,postgres"  # new: control which extras are installed
    validationRules: "all"           # new: enable all validation rules

• Build richer pre-built Docker images with broader provider coverage and popular data engineering packages

Contributions are very welcome. You can follow this standard contributions guidelines to contribute code.