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

Hello! I have a flow in 2.0 with two tasks that I want to run sequentially. I am using a docker flow runner and my flow looks something like this:

```@flow
def records_analysis_flow():
    ...
    future = ingest_csv(config)
    future.wait()
    future = analyze_records(ra_config)
    future.wait()```
I noticed somewhat inconsistent performance results. If I run the code in docker without prefect it takes around 154 seconds but when I run it in prefect it takes almost 900 seconds. However, if I explicitly set the task runner to sequential, then it takes ~ 180 seconds. Is my flow doing something wrong? Why does the concurrent task runner take so long? Thanks!

We’ll need more details about you flow/tasks, I’m surprised to see this behavior. If you can find out where the time is being spent, that’d be really helpful.

(or provide a minimal example that I can profile)

Is it still slow without those `wait` calls?

the first task is always super quick, the second one (which uses a language model) is the one that is really slow

but if I set the task runner to sequential it performs way better
```@flow(task_runner=SequentialTaskRunner())
def records_analysis_flow():
   ...```