Skip to main content

HAQM Managed Workflows for Apache Airflow
HAQM Managed Workflows for Apache Airflow is accessible in the next generation of HAQM SageMaker

HAQM Managed Workflows for Apache Airflow (MWAA)

Secure and highly available managed workflow orchestration for Apache Airflow

Get started with HAQM MWAA

Why HAQM MWAA?

HAQM MWAA is a managed service for Apache Airflow that lets you use your current, familiar Apache Airflow platform to orchestrate your workflows. You gain improved scalability, availability, and security without the operational burden of managing underlying infrastructure.

HAQM MWAA is accessible in the next generation of HAQM SageMaker

With HAQM MWAA in the next generation of HAQM SageMaker, you can deploy and scale Apache Airflow seamlessly without operational burdens. With automated scaling and built-in fault tolerance, MWAA in HAQM SageMaker ensures your workflows execute reliably—allowing you to focus on innovation, not infrastructure. Learn more. 

Missing alt text value

Benefits

Deploy Apache Airflow at scale without the operational burden of managing underlying infrastructure.

Run Apache Airflow workloads in your own isolated and secure cloud environment.

Monitor environments through HAQM CloudWatch integration to reduce operating costs and engineering overhead.

Connect to AWS, cloud, or on-premises resources through Apache Airflow providers or custom plugins.

HAQM MWAA powers workflows for the next generation of HAQM SageMaker with access to a personal, open-source Airflow deployment, running alongside Jupyter notebooks in HAQM SageMaker Unified Studio. You can easily develop Airflow Directed Acyclic Graphs (DAGs) that can orchestrate their project artifacts such as notebooks, queries, and training jobs.

Use cases

Create scheduled or on-demand workflows that prepare and process complicated data from big data providers.

Orchestrate multiple ETL processes that use diverse technologies within a complex ETL workflow.

Automate your pipeline to help machine learning (ML) modeling systems ingest and then train on data.