Hi, We are currently using a Azure AAD Token inorder to authenticate with Databricks instead of generating Personal Access Tokens from Databricks. We have a multi-tenant architecture and so we are using Azure container instances to run multiple transformation pipelines parallel using dbT.

Inorder to authenticate with databricks we generate an AAD token inside the ACI using the user defined managed identity (UDMI) which has contributor and reader rights on databricks. We do get the AAD token back successfully. However, the token when passed to databricks returns a 403 error.

UDMI is also an admin in the databricks workspace.

[0m18:11:15.979533 [debug] [MainThread]: Opening a new connection, currently in state init
[0m18:11:18.694997 [debug] [MainThread]: Databricks adapter: failed to connect: Error during request to server: : User not authorized.
[0m18:11:18.694997 [debug] [MainThread]: Databricks adapter: <class 'databricks.sql.exc.RequestError'>: Error during request to server: : User not authorized.
[0m18:11:18.694997 [debug] [MainThread]: Databricks adapter: attempt: 1/30
[0m18:11:18.695990 [debug] [MainThread]: Databricks adapter: bounded-retry-delay: None
[0m18:11:18.695990 [debug] [MainThread]: Databricks adapter: elapsed-seconds: 2.699450731277466/900.0
[0m18:11:18.695990 [debug] [MainThread]: Databricks adapter: error-message: : User not authorized.
[0m18:11:18.695990 [debug] [MainThread]: Databricks adapter: http-code: 403

Is there anything that we are missing and have to add more to this? Any help on this is appreciated.