hey community, I have a question that I've been unable to find an answer for in the docs. say I have two mutually exclusive flows, Flow A and Flow B. Flow B should not be executed until Flow A has completed successfully. how can I declare a dependency on Flow A from within Flow B? dagster has something similar to this and I was hoping that prefect does as well.

Hey @aaron, I think the paradigm here would be to use the

create_flow_run

task at the end of Flow A to trigger Flow B. You can also do it on the Flow B side, but it becomes harder because you have to poll for state using the GraphQL API. It will also take some thought how to point it to a specific Flow run of Flow A You can also have an outer main flow that is responsible for calling

create_flow_run

for both Flow A and Flow B.

thanks for the quick response! so there is no way to explicitly declare a dependency on an upstream task or flow by name, just triggering downstream? I realize my example just used two flows, and while your recommended approach would work for a few downstream dependents but if I have dozens of pipelines that are dependent on Flow A's successful execution, triggering them all from Flow A would get cumbersome.

But let’s say Flow B is checking if it can start, and Flow A has two copies, how does B know which copy of A it is dependent on?

yes, I believe so

I think the equivalent is a Flow of Flows in Prefect because they use the Graph to stitch together jobs and run them together. For us, the outer Graph would just be a “main” flow where you have all your flow logic. This is the docs page for that

and yes, just one concurrent run of Flow A. let's say this Flow refreshes a data source once daily. Other flows depend on this data source being up to date, but they should not run until Flow A has completed. Is the only way to accomplish this currently in Prefect to query the graph?

I think the flow of flows paradigm would be much easier. But yes we can also do it from the GraphQL API, but I think that’s fighting the framework honestly

yeah, I would have to agree. thanks for your help.

I am still reading about this Graph. The Graph definition seems pretty analogous to the main flow except they are more concerned with inputs and outputs. I think for you, you should just be aware that you can do the same with Parameters:

with Flow(...) as flow:
    x = Paramater("x", 1)
    flow_a = create_flow_run(...,parameters={"x": x})
    flow_b = create_flow_run(...,parameters={"x": x})