Forum Discussion

BharathKumarS's avatar
BharathKumarS
Regular Visitor
7 months ago
Solved

Can we use DefaultAzureCredential with Workspace Identity inside a Fabric Notebook?

Hi,    I’m trying to authenticate to Azure AD–protected resources (e.g., Microsoft Graph) from a Microsoft Fabric notebook using Workspace Identity, without secrets or certificates. My requirement...
notebook
  • Shreya_Barhate's avatar
    Shreya_Barhate
    7 months ago

    Hi BharathKumarS 

    Currently, using DefaultAzureCredential with Workspace Identity inside a Microsoft Fabric notebook is not supported. Workspace Identity is enabled at the workspace level, but it is not injected into the notebook runtime environment, which is why ManagedIdentityCredential fails with the IMDS endpoint error. This is a known limitation, and Microsoft has indicated plans to support identity injection for notebook runtimes in the future, though no official timeline has been announced. Until native support is available, the recommended approach is to use alternative authentication methods such as InteractiveBrowserCredential for user-based login or ClientSecretCredential with a service principal, storing secrets securely in Azure Key Vault or Fabric Key Vault. These methods provide a secure way to access Azure AD–protected resources like Microsoft Graph from Fabric notebooks.

    Thanks and regards,
    Shreya

deborshi_nag's avatar
deborshi_nag
Super User
7 months ago

Hi BharathKumarS 

 

No — today a Fabric notebook’s runtime does not expose the Workspace Identity to azure-identity, so DefaultAzureCredential() can’t succeed via ManagedIdentityCredential inside the notebook. That’s why you’re seeing “no response from the IMDS endpoint”—there is no managed‑identity endpoint wired into the Spark session. Workspace Identity is currently supported in Fabric connections/activities (pipelines, shortcuts, semantic models, Dataflows Gen2), not inside the notebook process itself.
 

Workspace Identity is used as an authentication method for connections used by OneLake shortcuts, pipelines (incl. Copy), semantic models, and Dataflows Gen2. 

 

Hope this helps! If it helps kindly indicate with a kudos or accepting as a solution.