Apache Airflow
Python-based workflow orchestrator — define pipelines as DAGs of tasks in code, schedule and visualise runs.
Maxime Beauchemin (Airbnb), 2014, donated to ASF 2016. The heart of modern data warehouses: extract job → transform job → load job, with retry, alerting and backfills baked in. Supports hundreds of operators (BashOperator, KubernetesPodOperator, SnowflakeOperator).
Criticised for Python-DAG magic (some idioms are unclear), scheduler overhead, heavy UI. Competitors: Dagster (typed, local-first), Prefect (modern Python API), Argo Workflows (Kubernetes-native). Airflow still wins on ecosystem breadth.