Skip to content
BENIBIX Learn

Amazon MWAA

Data🗓️ OrchestrationApache Airflow

Also known as Amazon Managed Workflows for Apache Airflow, MWAA, AWS Managed Airflow

AWS runs Apache Airflow for you. You bring the DAG files, in an S3 bucket.

Think of it like

It is like renting a car with a driver. You pick where to go, and someone else looks after the car.

Example

A team moves its DAGs from its own server to MWAA. It copies the files to an S3 bucket and stops patching Airflow by hand.

Why it matters

Running Airflow well takes real work: upgrades, scaling and security. MWAA hands that work to AWS.

Definition

Amazon MWAA, short for Managed Workflows for Apache Airflow, is where AWS runs Airflow for you. You choose an Airflow version, and AWS sets up the scheduler, the workers and the database. You put your DAG files in an Amazon S3 bucket, and MWAA picks them up. It adds more workers when there is more work. Logs and metrics go to Amazon CloudWatch. It is the same open source Airflow, so DAGs move in and out of it.
MWAA is one of the two highlighted boxes: the same Airflow, but AWS runs it.

Common mistake

Assuming every Airflow version is offered on the day it comes out. MWAA supports a set list of versions.

Key takeaways

  1. 1
    AWS runs it. Scheduler, workers and database.
  2. 2
    DAGs in S3. Upload files to a bucket.
  3. 3
    Real Airflow. The same open source code.