Forum Discussion
Gateway for Published Semantic Model in Mixed Mode
Hello all, we are just migrating to Powerbi and many of our users need information that we prepare in semantic models with build permission in our F64 Capacity. I've tried to search but I can't seem to find nor understand the documentation for this specific scenario and how to tackle it.
The scenario is this:
- I created a semantic model that connects to Snowflake and created a connection for it on powerbi service, it works fine and update without issues on the workspace.
- One of my users connected to that published semantic model and added an excel file, so the model became Mixed Mode and changed the connection to the published semantic model to Direct Query
- When the report is published, the new Mixed mode semantic model is also published, but now on settings is asking me to create a new connection for analysis services for my already published semantic model
I'm unable to understand why I need a new connection for a published semantic model, do I need to do this for every published semantic model? Is this a normal way to work?
thanks in advance
Hi maddypena07,
Using Workspace Identity to authenticate the new connection to the published semantic model is both correct and recommended, especially when the semantic model is in the same workspace and you want to avoid handling individual user credentials. Workspace Identity allows the service to use the workspace’s managed identity for access, which centralizes and secures permissions, making maintenance easier.
This process may seem complex at first because Power BI treats DirectQuery for Published Semantic Model similarly to a traditional Analysis Services connection. Even though everything is cloud-based, a “data source connection entry” is still needed for authentication validation and to enforce RBAC/security boundaries. That’s why this extra connection step is necessary when a Mixed Mode model references another semantic model.
Once the connection is set up and mapped properly, no gateway is needed as long as all sources are cloud-based. Refreshes and scheduled updates will work as expected without requiring a gateway.
This approach is better than having users create personal semantic models for each report, as it supports data governance, consistency, and reusable modeling across teams.
Thank you.
6 Replies
Edit: I created a connection but still ask me to use the gateway for some reason even though the connections are on cloud:
Hi maddypena07,
If you create a report using a published semantic model and add local data, such as from Excel, Power BI will automatically use DirectQuery for Published Semantic Model. This changes your model to Mixed Mode, making it behave as if it’s connecting to an Analysis Services source, even though it’s actually another Power BI semantic model in the cloud. As a result, the Service requires a data source connection entry for the published semantic model, similar to how it manages AS connections. This is why the Settings page requests an “Analysis Services” connection, even though everything is cloud-based. An on-premises gateway is only needed if you have a local source. Each mixed-mode dataset that connects to a published semantic model will need its own cloud data source connection entry to validate credentials for refreshes. This is standard for DirectQuery-on-Semantic-Model configurations. Once you set up the cloud connection with the same credentials as the original model, refreshes and connection checks should work without needing a gateway
Thank you.
Thank you for the detailed response!
I had to create a new connection to the published semantic model and use workspace identity as authentication, could be this the correct approach?
I find this behavior really convoluted, but I think is better than just let all users create their own personal semantic models for each report
Hi maddypena07,
Just wanted to follow up and confirm that everything has been going well on this. Please let me know if there’s anything from our end.
Please feel free to reach out Microsoft fabric community forum.Thankyou.
Hi maddypena07,
Using Workspace Identity to authenticate the new connection to the published semantic model is both correct and recommended, especially when the semantic model is in the same workspace and you want to avoid handling individual user credentials. Workspace Identity allows the service to use the workspace’s managed identity for access, which centralizes and secures permissions, making maintenance easier.
This process may seem complex at first because Power BI treats DirectQuery for Published Semantic Model similarly to a traditional Analysis Services connection. Even though everything is cloud-based, a “data source connection entry” is still needed for authentication validation and to enforce RBAC/security boundaries. That’s why this extra connection step is necessary when a Mixed Mode model references another semantic model.
Once the connection is set up and mapped properly, no gateway is needed as long as all sources are cloud-based. Refreshes and scheduled updates will work as expected without requiring a gateway.
This approach is better than having users create personal semantic models for each report, as it supports data governance, consistency, and reusable modeling across teams.
Thank you.
I'm so relieved to read your answer, thank you very much!
Yeah, I think that this scenario will present a lot in my organization, as the base model is provided by my team and complemented with user data, such as monthly objectives and so on.
Again, thank you for the detailed information, with this I'm also able to explain it to my team for future references!
Best day to you!
