Hi! I setup Kubernetes and I run a test project in AWS. Now I am updating the project locally, I run

python recommenderflow.py --datastore=local --metadata=local run

and the flow runs without problems locally. When I try to review the artifacts generated by the flow in a local jupyter notebook I get error as the client is trying to connect to AWS (none of the resources exist now after destroyed the env with terraform). How do I setup an environment variable to enforce using the local datastore and metadata when a kubernetes configuration is present?

There's "METAFLOW_DEFAULT_DATASTORE" and "METAFLOW_DEFAULT_METADATA" in the

~/.metaflowconfig/config.json

config which are probably getting loaded if you try to read data in the notebook

Thank you for the help Mikko!

As per the doc, in my notebook I run:

metadata('/path/to/flow_folder/')

and I was able to access the metadata