Hi I am trying to port our image processing pipeline to prefect. The analysis we run relays on a couple of configuration yaml file that are loaded from each process when the analysis runs in parallel. I was just wandering if I should keep this approach or add the parameters to the

prefect.context

and avoid the file loading. Is there any reason why I should not have the config data added to the context. Thanks a lot!

Hi @simone definitely both approaches are acceptable. Generally we encourage users to take advantage of things like Parameters if this is a parameterized workflow but if there is a big config that you want in context to be available to all tasks in the run then that is definitely acceptable as well 🙂

thanks!