Airflow 3.2 enables asset partitioning to trigger DAGs based on specific data slices, increasing accuracy in pipeline execution. It also enhances approval workflows with full audit trails, agentic operator HITL support, and synchronous callbacks for alerts. Visualization and task management improvements such as Grid View virtualization and full XCom UI control were included.
Available for new and upgraded MWAA environments, these features improve productivity in data engineering and orchestration at scale on AWS, facilitating finer granular scheduling and robust intervention controls in data workflows.
Source & further reading
Amazon MWAA now supports Apache Airflow 3.2 ↗Source published 2026-05-19. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.