Problem

When you create an MLflow experiment with a custom artifact location, you get the following warning:

Cause

MLflow experiment permissions (AWS | Azure | GCP) are enforced on artifacts in MLflow Tracking, enabling you to easily control access to datasets, models, and other files.

MLflow cannot guarantee the enforcement of access controls on artifacts stored in custom locations.

Solution

Databricks recommends using the default artifact location when creating an MLflow experiment.

The default storage location is backed by access controls.