Cross-Workspace Usage¶

By default, the advisor analyses a warehouse in the same workspace where your Fabric Spark notebook is running. To target a warehouse in a different workspace, you need to provide two additional identifiers.

Required Parameters¶

Both IDs are GUIDs that you can find directly in the browser URL when viewing the item. The URL format is:

https://app.fabric.microsoft.com/groups/{WorkspaceID}/warehouses/{WarehouseID}

Parameter	Where to Find
`workspace_id`	The GUID after `/groups/` in the URL when you are in the workspace
`warehouse_id`	The GUID after `/warehouses/` in the URL when the warehouse is open
`sql_endpoint_id`	The GUID after `/mirroredwarehouses/` in the URL when the SQL Analytics Endpoint is open

Choose warehouse_id or sql_endpoint_id

You must specify either warehouse_id or sql_endpoint_id, depending on where you want the advisor to run. Use warehouse_id to run the advisor against a Warehouse. Use sql_endpoint_id to run the advisor against a SQL Analytics Endpoint

Configuration¶

Both advisors accept the same cross-workspace parameters.

Data ClusteringPerformance Check

from fabric_warehouse_advisor import DataClusteringAdvisor, DataClusteringConfig

config = DataClusteringConfig(
    warehouse_name="TargetWarehouse",

    # Cross-workspace: both are required together
    workspace_id="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
    warehouse_id="yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy",
    # sql_endpoint_id="yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy",
)

advisor = DataClusteringAdvisor(spark, config)
result = advisor.run()

from fabric_warehouse_advisor import PerformanceCheckAdvisor, PerformanceCheckConfig

config = PerformanceCheckConfig(
    warehouse_name="TargetWarehouse",

    # Cross-workspace: both are required together
    workspace_id="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
    warehouse_id="yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy",
    # sql_endpoint_id="yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy",
)

advisor = PerformanceCheckAdvisor(spark, config)
result = advisor.run()

How It Works¶

Under the hood, the advisor passes these IDs as Spark reader options:

reader = spark.read \
    .option(Constants.DatabaseName, warehouse_name) \
    .option(Constants.WorkspaceId, workspace_id) \
    .option(Constants.DatawarehouseId, warehouse_id)

Where Constants comes from com.microsoft.spark.fabric.Constants — the Fabric Spark runtime's built-in connector.

Requirements¶

The identity running the notebook (your Entra ID / service principal) must have at least Read access on the target warehouse
The target warehouse / sql endpoint must be in the same Fabric tenant
Both workspace_id and either warehouse_id or sql_endpoint_id must be specified together — providing only one will not work

Same-Workspace Usage¶

When running in the same workspace, simply omit both parameters:

config = DataClusteringConfig(
    warehouse_name="MyWarehouse",
    # No workspace_id or warehouse_id needed
)

The Fabric Spark connector automatically resolves the warehouse in the current workspace.

Troubleshooting¶

Symptom	Cause	Solution
`RuntimeError: com.microsoft.spark.fabric is not available`	Running outside Fabric Spark	Must run in a Fabric Spark notebook
Permission error	Insufficient access on target warehouse	Grant Read access on the target warehouse to the notebook identity
Warehouse not found	Incorrect `warehouse_name` or `warehouse_id`	Double-check the name matches exactly (case-sensitive) and the GUID is correct
Empty results	Incorrect `workspace_id`	Verify the workspace GUID — it's the workspace containing the target warehouse, not the notebook's workspace