Invoking DAGs in different Amazon MWAA environments

The following code example creates an Apache Airflow CLI token. The code then uses a directed acyclic graph (DAG) in one Amazon MWAA environment to invoke a DAG in a different Amazon MWAA environment.

Version

• You can use the code example on this page with Apache Airflow v2 and above in Python 3.10.

Prerequisites

To use the code example on this page, you need the following:

• Two Amazon MWAA environments with public network web server access, including your current environment.

• A sample DAG uploaded to your target environment's Amazon Simple Storage Service (Amazon S3) bucket.

Permissions

To use the code example on this page, your environment's execution role must have permission to create an Apache Airflow CLI token. You can attach the AWS managed policy AmazonMWAAAirflowCliAccess to grant this permission.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "airflow:CreateCliToken"
            ],
            "Resource": "*"
        }
    ]

For more information, see Apache Airflow CLI policy: AmazonMWAAAirflowCliAccess.

Dependencies

• To use this code example with Apache Airflow v2, no additional dependencies are required. The code uses the Apache Airflow v2 base install on your environment.

Code example

The following code example assumes that you're using a DAG in your current environment to invoke a DAG in another environment.

1. In your terminal, navigate to the directory where your DAG code is stored. For example:

   cd dags

2. Copy the content of the following code example and save it locally as invoke_dag.py. Replace the following values with your information.

   • your-new-environment-name – The name of the other environment where you want to invoke the DAG.

   • your-target-dag-id – The ID of the DAG in the other environment that you want to invoke.

   from airflow.decorators import dag, task
   import boto3
   from datetime import datetime, timedelta
   import os, requests
   
   DAG_ID = os.path.basename(__file__).replace(".py", "")
   
   @task()
   def invoke_dag_task(**kwargs):
       client = boto3.client('mwaa')
       token = client.create_cli_token(Name='your-new-environment-name')
       url = f"https://{token['WebServerHostname']}/aws_mwaa/cli"
       body = 'dags trigger your-target-dag-id'
       headers = {
           'Authorization' : 'Bearer ' + token['CliToken'],
           'Content-Type': 'text/plain'
           }
       requests.post(url, data=body, headers=headers)
   
   @dag(
       dag_id=DAG_ID,
       schedule_interval=None,
       start_date=datetime(2022, 1, 1),
       dagrun_timeout=timedelta(minutes=60),
       catchup=False
       )
   def invoke_dag():
       t = invoke_dag_task()
   
   invoke_dag_test = invoke_dag()

3. Run the following AWS CLI command to copy the DAG to your environment's bucket, then trigger the DAG using the Apache Airflow UI.

   $ aws s3 cp your-dag.py s3://your-environment-bucket/dags/

4. If the DAG runs successfully, you'll see output similar to the following in the task logs for invoke_dag_task.

   [2022-01-01, 12:00:00 PDT] {{python.py:152}} INFO - Done. Returned value was: None
   [2022-01-01, 12:00:00 PDT] {{taskinstance.py:1280}} INFO - Marking task as SUCCESS. dag_id=invoke_dag, task_id=invoke_dag_task, execution_date=20220101T120000, start_date=20220101T120000, end_date=20220101T120000
   [2022-01-01, 12:00:00 PDT] {{local_task_job.py:154}} INFO - Task exited with return code 0
   [2022-01-01, 12:00:00 PDT] {{local_task_job.py:264}} INFO - 0 downstream tasks scheduled from follow-on schedule check

   To verify that your DAG was successfully invoked, navigate to the Apache Airflow UI for your new environment, then do the following:

   1. On the DAGs page, locate your new target DAG in the list of DAGs.

   2. Under Last Run, check the timestamp for the latest DAG run. This timestamp should closely match the latest timestamp for invoke_dag in your other environment.

   3. Under Recent Tasks, check that the last run was successful.