Managed Apache Airflow
What's different
Google's offering (renamed from Cloud Composer in 2026) is the most mature. Azure Data Factory's managed Airflow stopped accepting new instances in January 2026; Microsoft points to Fabric.
Analytics & Data · Managed Apache Airflow
Managed Apache Airflow, side by side on AWS, Azure and Google Cloud. They do the same job, but differ in features, limits or pricing, so check the notes below before migrating.
Managed Apache Airflow
Google's offering (renamed from Cloud Composer in 2026) is the most mature. Azure Data Factory's managed Airflow stopped accepting new instances in January 2026; Microsoft points to Fabric.
| Aspect | AWS | Azure | Google Cloud |
|---|---|---|---|
| Service | Amazon MWAA | Apache Airflow jobs in Microsoft Fabric | Managed Service for Apache Airflow (renamed from Cloud Composer in 2026) |
| Maturity | Established, widely used | Newer; replaces Data Factory's managed Airflow | Most mature; the current generation (Gen 3) supports Airflow 3 |
| How it's billed | Per environment size plus workers | Through Fabric capacity | By the environment's compute and storage |
| Same DAGs? | Standard Airflow DAGs | Standard Airflow DAGs (copied over, not converted) | Standard Airflow DAGs |
Your data and pipelines live on AWS.
You're adopting Microsoft Fabric. Check networking needs first, since Fabric's Airflow has fewer private-network options.
You want the most mature managed Airflow, or your data is in BigQuery.
Details for Amazon MWAA verified November 2026 against official documentation: Azure: Workflow Orchestration Manager (archived) · Amazon MWAA · Google: Managed Service for Apache Airflow.