Skip to main content
Last updated on

Troubleshoot publishing to Power BI

This page lists common errors you might encounter when connecting to or publishing Databricks tables to Power BI, along with how to resolve them.

"Approval required" when you first launch Publish to Power BI​

You need Entra Admin approval before you can grant permissions to this integration feature. Contact your Entra Admin to approve the request. For instructions, see Review admin consent requests.

No permission to call the Discover method​

The '<user>' user does not have permission to call the Discover method

Ensure that you are publishing to a Power BI workspace with a Premium license mode. You cannot publish to a workspace with a Pro license mode.

No permission to create a new object​

The '<user>' user does not have permission to create a new object

You might have the Viewer role in the Power BI workspace. Check whether you have the permissions to create a semantic model. For more information, see Roles in workspaces.

PowerBINotLicensedException​

Ensure that you have a Power BI Premium license (Premium capacity, Premium Per User (PPU), or Microsoft Fabric capacity).

XMLA read/write support is disabled​

You might see one of the following errors:

We couldn't deploy this dataset. Contact a capacity administrator to make sure XMLA read/write support is enabled in the capacity settings on the Power BI Premium capacity, and then try again. For additional information, see 'XMLA read/write support' in the product documentation.

XMLA endpoint feature is disabled. Turn on Analyze in Excel feature in PowerBI.com to enable this feature.

Contact your Power BI capacity administrator to enable Read and Write capabilities under XMLA endpoint in the capacity settings. For instructions, see Enable XMLA read-write.

CapacityNotActive​

Your Fabric or Power BI capacity might be paused. Contact your capacity administrator to check the capacity status.

"Failed to update data source credentials" when editing credentials​

When editing data source credentials in Power BI, you might see:

Failed to update data source credentials: [Microsoft][thriftextension] (14) Unexpected response from server during a HTTP connection: Unauthorized/Forbidden error response returned, but no token expired message received.

Check whether your Databricks workspace is publicly accessible. If your workspace uses Private Link or IP access lists, then you might need to configure a Power BI on-premises gateway.

Semantic model name already exists​

The database with the name of '<name>' already exists

Ensure that you have the permissions to create a semantic model. Then, retry Publish to Power BI to publish the semantic model with an incremented version number.

Direct Lake mode conflict​

You cannot use Direct Lake mode together with other storage modes in the same model

Since Publish to Power BI creates a semantic model with Import or DirectQuery mode, it cannot be used to publish to an existing dataset that uses Direct Lake mode. To learn more, see Direct Lake overview.

Semantic model stuck on "Pending changes"​

After a Power BI task or Publish to Power BI updates a semantic model, the model is stuck on Pending changes with no way to apply or discard them.

This is a Power BI platform limitation, not a Databricks issue: Power BI doesn't process external XMLA metadata writes while the model is open in an edit session, so the editor session becomes stale. There's no API to detect whether someone has the model open in edit mode before publishing.

To clear the state, connect to the model through the XMLA endpoint (for example, with Tabular Editor) and apply or discard the pending changes. To avoid the issue, don't keep a semantic model open for editing while a Power BI task or Publish to Power BI job runs. If the issue persists, contact Microsoft support.

On this page