Hey guys, had a quick question regarding workspace issues when using dagster-docker: when I was using dagster with a non-dockerized deployment, code issues that led to problems with the repository being loaded (for ex. with imports) would typically pop up when attempting to deploy, but now that I am using a dockerized dagster deployment if a similar issue arises I just see some variant of "dagster.core.errors.DagsterUserCodeUnreachableError: Could not reach user code server". Is there a good way to see what the underlying issue is for cases like this?

I usually use

docker logs <container_name>

or

docker compose logs <container_name>

(where container name is the name of the user code container) to debug these kinds of issues - it should give you the actual stack trace from the user code container. You can get the container name by listing the containers with

docker ps

or

docker compose ps

Thanks Zach, that was what I ultimately ended up doing as well, though it would be nice to be able to access that through dagster

I agree

@Dagster Bot issue Better support for debugging local container issues for user code deployment errors

Created issue at: https://github.com/dagster-io/dagster/issues/8741

Hey Carter - there's also a --lazy-load-user-code flag that you can use when launching your grpc server: https://docs.dagster.io/_apidocs/cli#cmdoption-dagster-api-grpc-lazy-load-user-code If you have that set, and there's an error while loading your code, it will leave the server running and display the error in dagit instead of making it shut down