I have a number of flows (50+) that are all lightweight in and of themselves, and all have similar (practically identical) dependancies. We use Docker for storage, so, it seemed to make the most sense to bundle up all flows into a single Docker container. So far this scheme has been ideal for our use case. However, if one flow is updated, all flows need to be reassigned to the same storage and reregistered (right?). This is fine, except it is very annoying that in the UI all flows increment in version. If I make 15 changes across 50 flows, a flow that was  never  edited since its initial deployment will be on version 16 😕 Is there anyway to preferentially prevent version incrementing, or any other suggests for how to be approaching this, besides having a different container for each flow?

For one, we are moving to user-defined versions in the very near future to avoid issues like this. I think you may need to just build that container separately with DockerRun and then use some other storage like S3 or Github?

Or you do use the CLI to register?

What do you mean build a container separately with DockerRun and then use some other storage like S3 or Github?

If you have all of them in the same script, I think it's unavoidable because that

register()

call will surely up the version. I think you need them split out into multiple files, but then you run into a problem of knowing when you need to rebuild the container

I'm just throwing potential ideas, I don't have the concrete solution ironed out

and so you know, I actually dont have all flows in the same script, because it didnt make sense to me organizationally. Each flow is in its own file, but I have a separate script I call

flow_registry

that relies on the

extract_flow_from_file()

utility to crawl my

flows/

directory, import all flows, and then loop through them assigning storage and registering with Prefect cloud