mlflow_experiment.md

subcategory
MLflow

databricks_mlflow_experiment Resource

This resource allows you to manage MLflow experiments in Databricks.

Example Usage

data "databricks_current_user" "me" {}

resource "databricks_mlflow_experiment" "this" {
  name              = "${data.databricks_current_user.me.home}/Sample"
  artifact_location = "dbfs:/tmp/my-experiment"
  description       = "My MLflow experiment description"
}

Argument Reference

The following arguments are supported:

• name - (Required) Name of MLflow experiment. It must be an absolute path within the Databricks workspace, e.g. /Users/<some-username>/my-experiment. For more information about changes to experiment naming conventions, see mlflow docs.
• artifact_location - Path to dbfs:/ or s3:// artifact location of the MLflow experiment.
• description - The description of the MLflow experiment.

Access Control

• databricks_permissions can control which groups or individual users can Read, Edit, or Manage individual experiments.

Import

The experiment resource can be imported using the id of the experiment

$ terraform import databricks_mlflow_experiment.this <experiment-id>

Related Resources

The following resources are often used in the same context:

• End to end workspace management guide.
• databricks_directory to manage directories in Databricks Workpace.
• databricks_mlflow_model to create MLflow models in Databricks.
• databricks_notebook to manage Databricks Notebooks.
• databricks_notebook data to export a notebook from Databricks Workspace.
• databricks_repo to manage Databricks Repos.