Working with Apache Airflow™ interfaces

To work with Managed Service for Apache Airflow™, you can use the web interface or the Apache Airflow™ API.

Only users with the managed-airflow.user role or higher can open the web interface or send an API request. The API checks the role via the IAM token used for authentication.

Apache Airflow™ interfaces can be accessed from the internet, so the role check provides additional security. Even if an attacker has a link to the web interface and the admin password to get authenticated with, they will not have the role required to access the interface.

Working with the web interfaceWorking with the web interface

1. In the management console, open the cluster page.

2. Go to the Managed Service for Apache Airflow™ web server page.

3. In the web interface that opens, enter the credentials:

   • Username: Admin
   • Password: Admin user password that you set when creating a cluster

For more information about how to use the web interface, see the Apache Airflow™ documentation.

Using the Apache Airflow™ APIUsing the Apache Airflow™ API

You can make Apache Airflow™ API requests from VMs located in your Managed Service for Apache Airflow™ cluster cloud network.

1. In the cloud network hosting the Managed Service for Apache Airflow™ cluster, create a Linux-based VM.

2. Set up a NAT gateway to connect to the VM.

3. Get an IAM token for a user account with the managed-airflow.user role or higher. The way you get a token depends on your account type:

   • Yandex account
   • Service account
   • Federated account

4. Make an API request, e.g., using cURL. In the authorization parameters, enter your username and password.

   Request example:

   curl \
       --request GET \
       'https://c-<cluster_ID>.airflow.yandexcloud.net/api/v1/dags' \
       --user 'admin:<admin_user_password>' \
       --header 'Content-Type: application/json' \
       --header 'X-Cloud-Authorization: Bearer <IAM_token>'
   

   The request returns a list of DAG files. Check an example of the response in the Apache Airflow™ API documentation.