The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Hello. I have *job A* that can run large numbers of runs in parallel, each run creates an asset. I have an asset sensor running every second - the idea is each asset created by *job A* triggers a run of *job B*.

Unfortunately the asset sensor misses some materialisation (when two *job A* runs finish at the same time). How can I make it so that:
```number of A runs == number of B runs```

For instance, is it possible to yield a `RunRequest` from inside an op?

It seems like this is not possible, I will refactor my pipeline to combine both jobs into one.

Hi Archie, Are you still having problems with this? If so, by:

&gt; each asset created by job A
Do you mean each materialization of a single asset?

Maybe these two can help:
<https://docs.dagster.io/concepts/partitions-schedules-sensors/asset-sensors#asset-reconciliation-sensor>
<https://docs.dagster.io/concepts/partitions-schedules-sensors/asset-sensors#triggering-runs-upon-materializations>

The second link is specifically for multi-assets but you can probably re-use that code to ensure that unconsumed materializations will trigger jobs

Thanks for the help all! But in the end it made sense to use a single job and multiple graphs for my architecture :slightly_smiling_face: