Issues with Mixed Authentication Methods in Dataflows Gen2

[Udo_S]

Hello everyone,

We are using several Dataflows Gen2 with mixed authentication methods. For example, we have a Dataverse connector using Organizational Account authentication, and an on-premises SQL connector (via local gateway) using Basic authentication — both within the same Dataflow.

This setup worked fine until a few weeks ago, when the dataflows began to fail with a rather cryptic error message. After some testing, we discovered that using mixed authentication methods in a single dataflow no longer seems to work.

I’m aware that the current workaround is to split the data sources into separate dataflows — one for each authentication method — and then join the results. This is how we're handling it for now.

Nevertheless, I wanted to reach out to check whether there's another solution, or perhaps a hidden checkbox deep within Fabric that could allow mixed authentication again.

Thanks in advance!

Udo

[Udo_S]

Hi @Anonymous ,
Thanks a lot for the clarification. Fortunately, there is no major impact - it's just a little inconvinient to keep everything separated.
Best regards,
Udo

[Anonymous · Accepted solution]

Hi @Udo_S,

Thank you for reaching out to the Microsoft Fabric Forum Community.

as of recent changes in Microsoft Fabric / Power BI, mixed authentication methods within a single Dataflow Gen2 are no longer supported, even though this used to work previously.

Keep using the split dataflows method that’s the workaround at this point. If your organization is heavily impacted, you may want to submit a support ticket or vote on an idea in the Power BI Ideas forum, especially if this regression wasn’t well documented. Please refer the below document to get more details.

https://learn.microsoft.com/en-us/fabric/data-factory/connector-azure-data-lake-storage-gen2

Best regards,
Prasanna Kumar