Connect GX Cloud and Airflow

In this quickstart, you'll learn how to use GX Cloud with Apache Airflow directly. You'll create a basic DAG that runs a Checkpoint in GX Cloud, and then trigger it through a local installation of an Airflow server.

Try the Great Expectations Airflow Provider

For a more ergonomic interface and more flexibility, use the Great Expectations Airflow Provider maintained by Astronomer to validate data directly from a DAG.

Apache Airflow is an orchestration tool that allows you to schedule and monitor your data pipelines. For more information about Apache Airflow, see the Apache Airflow documentation.

Prerequisites

• You have a GX Cloud account.

• You have installed Apache Airflow and initialized the database (airflow db init).

• You have connected GX Cloud to a Data Asset on a Data Source. (Note that this automatically creates the Checkpoint your DAG will run.)

• You have added Expectations.

Run Airflow Standalone to create a fresh local Airflow environment

1. The airflow standalone command initializes the database, creates a user, and starts all components.

   Terminal input

   airflow standalone

   This command will eventually output a username and password for the Airflow UI like this:

   Terminal input

   standalone | Airflow is ready
   standalone | Login with username: admin  password: Bpu6RgmPMMaDeeq5
   standalone | Airflow Standalone is for development purposes only. Do not use this in production!

2. Access Airflow UI:

   Once the web server is running, open a web browser and go to http://localhost:8080 (by default) to access the Airflow UI using the username and password from the last step

Create a DAG file for your GX Cloud Checkpoint

1. Open a terminal, browse to the airflow folder in your home directory, and then run the following code to create a new DAG named gx_dag.py:

   Terminal input

   cd ~/airflow
   mkdir dags
   cd dags
   touch gx_dag.py

2. Open the gx_dag.py DAG file and add the following code:

   gx_dag.py

   import os
   import pendulum
   import great_expectations as gx
   from airflow.decorators import dag, task
   
   @dag(
       schedule=None,
       start_date=pendulum.datetime(2024, 9, 1),
       catchup=False,
   )
   def gx_dag_with_deco():
       os.environ["NO_PROXY"] = "*" #https://github.com/apache/airflow/discussions/24463
       print("Great Expectations DAG Started")
   
       @task
       def run_checkpoint():
           print("Running Checkpoint")
           # Replace <YOUR_ACCESS_TOKEN>, <YOUR_CLOUD_ORGANIZATION_ID> with your credentials
           # You can also set GX_CLOUD_ACCESS_TOKEN and GX_CLOUD_ORGANIZATION_ID as environment variables
           GX_CLOUD_ACCESS_TOKEN = "<YOUR_ACCESS_TOKEN>"
           GX_CLOUD_ORGANIZATION_ID = "<YOUR_CLOUD_ORGANIZATION_ID>"
           # Find the Checkpoint name in the GX Cloud UI. 
           # - Go to the "Validations" tab.
           # - Next to the "Validate" button, click the code snippet icon.
           # - Click "Generate snippet". 
           # - Copy the Checkpoint name from the code snippet and use it below.
           CHECKPOINT_NAME = "my_data_asset 123ABC - Default Checkpoint"
           context = gx.get_context(
               mode="cloud", 
               cloud_organization_id=GX_CLOUD_ACCESS_TOKEN, 
               cloud_access_token=GX_CLOUD_ORGANIZATION_ID
           )
           checkpoint = context.checkpoints.get(CHECKPOINT_NAME)
           checkpoint.run()
           return f"Checkpoint ran: {CHECKPOINT_NAME}"
       run_checkpoint()
   
   run_this = gx_dag_with_deco()

3. Save your changes and close the gx_dag.py DAG file.

Run the DAG (Manually)

1. Restart the airflow server to pick up the new DAG file.

2. Sign in to Airflow using the username and password from the first standalone run

3. In the Actions column, click Trigger DAG for gx_dag and confirm your DAG runs as expected.

Clean up local Airflow environment

1. Delete the local files and sqllite database

   Terminal input

   rm -rf ~/airflow