Managed Apache Airflow: Scaling Data & AI Workloads

Google Cloud has rebranded Cloud Composer as Managed Service for Apache Airflow and announced four major updates. Apache Airflow 3.1 is now generally available, bringing decoupled architecture, DAG versioning, managed backfills, event-driven scheduling, and Human-in-the-Loop support. A Data Engineering Agent powered by Gemini is embedded in the dashboard for agentic troubleshooting of DAG failures. A new declarative orchestration framework lets users define pipelines in YAML without deep Airflow expertise, integrating with dbt, Spark, and other tools via GitHub Actions. Finally, an MCP Server for Managed Airflow is in public preview, exposing tools for agents and developers to query environment and DAG run data directly from their preferred interfaces.