HI team, I am looking to build inter-job dependencies and trigger downstream jobs when an upstream job succeeds. It seems that `run_status_sensor`() is the way to go. I see that it support monitored_jobs and monitored_pipelines. For monitored_jobs, could I reference a job by its string name instead of the job object?

This isn't currently possible, but we might be able to make it possible soon. Out of curiosity, what's the reason you'd prefer to use a string than a job object? Here's a related request that we got recently: https://github.com/dagster-io/dagster/issues/9394

there are a few other ways to do this, but nothing exactly like what you're saying. For example, it's possible to trigger a job using a sensor if you create a new object in a database. Dagster also has an “asset sensor” which does something similar. Main limitation is that it does not trigger a job for each new asset materialization, just the latest one (although there is an open MR to fix that here: https://github.com/dagster-io/dagster/issues/5699)

Thanks @sandy. I see the

run_status_sensor()

monitored_jobs param could take in an

UnresolvedAssetJobDefinition

. I am wondering if I could use

UnresolvedAssetJobDefinition

with an asset job name (or something?) to reference another job? (Context: we dynamically generate jobs based on yamls defined by our users. So when we are generating each job, we don’t have the complete references to all jobs.)

To be completely honest I’m not sure. In my use case the cursor wasn’t useful, so I didn't look into it.

I see. Thanks

@Hebo Yang it's possible that your

UnresolvedAssetJobDefinition

scheme would work, but you'd need to try it out to be sure. when this change is merged, we should have a more supported way of doing this: https://github.com/dagster-io/dagster/pull/9405

Got it. Thanks Sandy!