Hi! Is there any way to set a data retention policy for runs? Not interested in storing old run history / logs. Documentation only mentions schedule / sensor data.

Hi David. Unfortunately there is no easy way to do this automatically. One thing that you could try is set up a job on a schedule that goes back and deletes old runs, using the instance methods (e.g.

context.instance.delete_run(run_id)

). By deleting runs, you might also delete some associated metadata (e.g. asset materializations), not sure if that matters to you or not

Hi Claire, I think this could work for us! Thanks!!!