The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

I have a question about `dagster-mlflow` plugin.
Mlflow version is pinned to 1.26.0 in the code already (<https://github.com/dagster-io/dagster/blob/master/python_modules/libraries/dagster-mlflow/setup.py#L34>)
Could you remove mlflow package from setup.py and give developers a decision of which package version they want to install?
I would like to use mlflow==1.26.1 version becasue it is still working fine with `mlflow sagemaker build-and-push-container` command

Is there any workaround or trick to use modern/latest Mlflow version?

Hi Marcin. Unfortunately our integration tests still break with mlflow==1.26.1. In order for us to release this pin I think we'd also have to upgrade our protobuf dependency, which is in progress: <https://github.com/dagster-io/dagster/issues/10627>

For now, I'd recommend filing an issue for this

Thx for the feedback.   I am using `protobuf==3.19.5` in the project. Mlflow and Dagster are accepted this version but it is true that I could hit a bug with gRPC issue with this version on my host.
I will wait for the fix. Thank for you help Claire and team