Unable to connect When Using Azure Databricks source from Power BI Desktop

[ManasRout]

Hi Team, While connect data from Azure databricks from power bi desktop getting below error, I have checked using correct hostname, path, latest version of power bi desktop also cluser is running still not able to connect. Is it blocking any Zscaler or something.

[ManasRout]

Hi @SharmilaBrisca Thank you for sharing these details I have checked with this user he can able to use databricks native connector without any issue also in this workspace only he is getting this issue rest all can able to connect.

[v-karpurapud]

Hi @ManasRout 

Thanks for the update. Since the same user can connect to other Databricks workspaces and other users can connect to this workspace, we can rule out network, proxy, or general connectivity issues. This suggests the problem is related to user-specific configuration.
 

The error (ADBC 400 – Bad Request) means the request from Power BI is reaching Databricks, but it's not valid for the endpoint. This usually happens when the connection details don't exactly match the compute resource.
 

To fix this, go to the Databricks workspace and open the Cluster or SQL Warehouse you want to connect to. Use the “Connect → Power BI” option to copy the Server Hostname and HTTP Path from the workspace, making sure they match the correct compute type.

SQL Warehouses should use  /sql/1.0/endpoints/<endpoint-id>, and
clusters should use   /sql/protocolv1/o/<org-id>/<cluster-id>.
 

Since it’s only affecting one user, check their configuration as well. Have them generate a new Personal Access Token (PAT) for authentication, clear any cached credentials in Power BI Desktop under Data Source Settings, and make sure they have the right permissions on the SQL Warehouse (“Can Use”) or cluster (“Can Attach To”).

I hope this clarifies things. Let me know what you find after trying these steps; I'm happy to assist further if needed.

Regards,

Microsoft Fabric Community Support Team.

[ManasRout]

Hi @v-karpurapud No this user can able to connect databricks connector and able to connect from power bi desktop but when usisng Azure Databricks getting the above error. In this workspace only this user is facing issue while using Azure Databricks other users can able to connect both databricks native as well as Azure Databricks. So I have created a ticket to our Zscaler team let's see if they can able to help this user.

[v-karpurapud]

Hi @ManasRout 

Thank you for the update. Hope your issue gets resolved soon. when it does, please share the insights with the community as it can be helpful to others.

Regards,

Microsoft Fabric Community Support Team.

[ManasRout · Accepted solution]

Hi @v-karpurapud , @SharmilaBrisca the issue got resolved after cleared cached and remove data source permission then user can able to use azure databricks connector from databricks to power bi.

[v-karpurapud]

Hi @ManasRout 

Glad to hear the issue has been resolved. If you have any other questions, feel free to contact us. We're here to help.

Regards,
Microsoft Fabric Community Support Team.

[SharmilaBrisca]

Hi @ManasRout ,

That error isn’t random-HTTP 400 (Bad Request) from ADBC usually means Power BI is reaching Databricks, but something in the connection details or network layer is invalid or being altered. Zscaler can cause it, but it’s not the only reason.

Here’s a step-by-step troubleshooting path that works in real projects:

1. Validate Databricks Connection Details (Most common issue)

Double-check these exactly from Azure Databricks workspace → Advanced Options:

• Server Hostname
• Format:

adb-<workspace-id>.<region>.azuredatabricks.net

• HTTP Path
• For cluster:

/sql/protocolv1/o/<org-id>/<cluster-id>

• For SQL Warehouse:

/sql/1.0/endpoints/<endpoint-id>

 Mistakes here = 400 error immediately

1. Authentication Check (Critical)

Use Personal Access Token (PAT):

• Go to Databricks → User Settings → Access Tokens
• Generate new token (don’t reuse old/expired)

In Microsoft Power BI Desktop:

• Authentication method → Basic
• Username: token
• Password: <your PAT>

 Wrong auth method = 400 error

1. 3. Zscaler / Firewall / Proxy Issue (Very likely in your case)

Yes — Zscaler can break ADBC (Thrift over HTTPS).

Symptoms:

• 400 error
• Works in browser but not in Power BI
• Happens only in corporate network

Fix:

Ask your network/security team to allowlist:

• *.azuredatabricks.net
• *.cloud.databricks.com

And bypass SSL inspection for:

• Port 443
• Thrift/HTTPS traffic

 Zscaler SSL inspection often corrupts ADBC requests → causes 400

1. 4. Connector Version / Driver Issue

Ensure:

• Latest Power BI Desktop (Monthly) (not Store version)
• Use Azure Databricks connector (native)

If still failing:

• Try ODBC Driver (Simba Spark ODBC)

1. 5. Cluster / SQL Warehouse State

Even if “running”, check:

• Cluster mode = Standard / High Concurrency
• Not Single Node (sometimes restricted)
• Permissions:
• You have Can Attach To
• You have Can Use SQL Endpoint

1. 6. Test Outside Power BI (Isolate issue)

Try connecting using:

• Databricks SQL Editor
• Or ODBC test

If that works → issue is Power BI / network layer

1. 7. Known Root Causes for THIS exact error

This specific error:

Thrift.Transport.Client.THttpTransport
Response status code: 400 (Bad Request)

Typically caused by:

• ❌Wrong HTTP Path
• ❌Expired PAT
• ❌Proxy rewriting headers (Zscaler)
• ❌SSL inspection breaking Thrift protocol

Please refer the below Links more information:

• Azure Databricks + Power BI:
  https://learn.microsoft.com/en-us/azure/databricks/partners/bi/power-bi
• Databricks SQL connectivity:
  https://learn.microsoft.com/en-us/azure/databricks/integrations/bi/power-bi
• Network requirements:
  https://learn.microsoft.com/en-us/azure/databricks/resources/ip-domain-region
• Power BI troubleshooting:
  https://learn.microsoft.com/en-us/power-bi/connect-data/service-azure-databricks
  
  Hope this resolve your problem

Thanks and Regards
GainInsights Solutions
Trusted Microsoft Partner - https://gain-insights.com/partnerships/microsoft/
www.gain-insights.com