Airflow is not going anywhere, and neither is our Airflow provider. But not every data workflow is a DAG, and forcing event-driven work into DAG primitives has a cost we have watched customers pay.
Two execution models, one catalog
You should pick the model based on what actually triggers the work: if the trigger is a clock, you want Airflow. If the trigger is data — a new commit on a branch, a partition crossing a threshold — you want an event subscription.
PV
Written by
Priya Venkatesan
Head of Product at DXData.