Hey guys, is it possible to launch multiple individual runs with the

asset_reconciliation_sensor

in oppose to a single run materializing all the assets? My only current resource management option is the

QueuedRunCoordinator

which I'm using to limit concurrency for my assets / jobs based on tags. As I understand, this will no longer work with

asset_reconciliation_sensor

. Maybe @owen can help?

hi @Daniel Gafni! the quick answer to the initial question is that this is not currently possible (the sensor will group together executions into a single logical run where possible), although the reason for this is mostly that it makes it easier for humans to see/understand what the sensor is doing, rather than any fundamental technical reason. As we make improvements to the UI/UX of this sensor, it's possible that this might change (it still can be hard to reason about what the sensor is doing!). However, just so I'm understanding your ideal setup, were you originally targeting a single asset per job, and therefore had an overall "no more than N assets should be executing at any given point in time" constraint? Or were there some jobs with multiple assets that could be running at the same time, or different tags for different kinds of compute?

Yes, I originally had a single job per asset