We have ~8 flows all using Docker storage and KubernetesJobEnvironment with yaml files.

We’d like, I think, to use run_configs to get rid of the yaml files, github storage to be able to update our flows without a giant image build step, and LocalDaskExecutor to parallelize our mapped tasks.

run_configs seems like a pretty straight forward change, but I’m not sure I understand the relationship between the containers, executor, and file based storage. I’ve spent a bunch of time in the docs, but can’t seem to put the pieces together. It seems like we would still need to build a container per flow, then tell the container to pull the flow file from GitHub. There also seem to be some changes that would require a rebuild of the image, but it’s not entirely clear what they would be outside of changes in Python requirements. I’m also not sure how a LocalDaskExecutor plays with a K8s deployment. I’m guessing it just runs the local cluster on the container?