Historically Airflow was only capable of time-based scheduling, where a DAG would run at certain times. For data updates at varying times, such as an external party delivering data to an S3 bucket, that meant having to run a DAG and continuously poll for updates. Airflow 3 introduces event-driven scheduling that enables you to trigger DAGs based on such updates. In this talk I'll demonstrate how this changes your DAG's code and how this works internally in Airflow. Lastly, I'll demonstrate a practical use case that leverages Airflow 3's event-driven scheduling.