Airflow is a tool to programmatically create, schedule and monitor data pipelines. By treating Directed Acyclic Graphs (DAGs) as code, it encourages maintainable, versionable and testable data pipelines. We've leveraged this configuration in our projects to create dynamic pipelines that resulted in lean and explicit data workflows. Airflow makes it easy to define your operators and executors and to extend the library so that it fits the level of abstraction that suits your environment.