Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Can you run another independent deployment when a flow run finishes? I have 2 flows that use different pools of workers, one is cpu heavy and another is gpu heavy. I want to  run the first one on my cpu nodes, and if they were successful, run the second flow on a different worker.

Why not just use the `run_deployment`  method?

U can call prefect to run a deployment from within a flow

This is doable (by doing this in `on_completion` ) of the first flow.

But this tightly couples my flows together.

I have to pass parameters to my first flow to give to the second flow.

<https://github.com/PrefectHQ/prefect/blob/d1d0e175a44c5d442936ed34fd2f2fb32ddda06b/src/prefect/deployments/deployments.py#L50>

I have used `run_deployment` extensively in my project. The problem is that calling the orchestrator in a flow to run another flow seems like the least viable option. It's both a security risk (my flows are running client code) and also I have to tightly couple the flows together, so `flow_a` has to have some unneeded information that only `flow_b` needs to have to be executable.

If I could chain multiple deployments together conditionally making DAGS would have been a lot easier.

Write `flow_c`  that takes input of both `flow_a`  and `flow_b` . :smile:

:smile:
The main problem is that the infrastructure for both flows are separate and I have to run them sequentially.

U can deploy flow1 on a queue that is queied by agent with high CPU.
Deploy flow2 on a queune that isqueried by agent with GPU.

At the end of flow1 `run_deployment` , deployment of flow2.