Skip to content

Connections

The Connections page provides all the credentials and endpoints required to connect to warehouse services from external applications. Use the drop-down menu to switch between connection types.

For ML workloads, you need the following three values:

Value Description
Spark Connect host The endpoint for connecting to the Spark Connect Server from an external client.
JWT bearer token The authentication token required by all external service endpoints.
MLflow Tracking URI The endpoint for logging and retrieving ML experiment data by using the MLflow client.
MLflow artifact location The bucket path that will contain persisted models.

To connect to the Spark server from a local machine, retrieve the host address and authentication token from the Connections page.