Proxy Model are not supported

Meaning that it contains native queries (Power Query steps that use a “native query” pattern, typically via the connector’s SQL/ODBC query. Deployment pipelines currently don’t support deploying proxy semantic models that include native queries, so the pipeline blocks the move. You ronly option is to remove native queries from the imported part.

https://learn.microsoft.com/en-us/fabric/cicd/deployment-pipelines/understand-the-deployment-process?tabs=new-ui 

Hey TaroGulati  ,

The error you’re seeing:

“proxy model semantic models which include native queries are not supported”

comes from a limitation in Power BI deployment pipelines — not from a mistake in your model structure.

What’s happening

When you build a model that mixes:

Import tables (your 5 imported tables)

DirectQuery connection to another Power BI dataset/semantic model

Power BI creates a proxy/semantic model behind the scenes. Deployment pipelines currently do not support proxy models that include native queries or DirectQuery behavior together which includes your scenario where one part of the model is DirectQuery to another published dataset and the other is import.

Why the error appears

Deployment pipelines treat composite models that connect to other Power BI datasets as “proxy models.”

If there’s any native query involvement or mixed connectivity mode like this, pipelines block it.

This is by design in the Power BI/Fabric deployment pipelines system it isn’t something you’re doing incorrectly.

You can see Microsoft’s official note about semantic model limitations for pipelines here:
https://learn.microsoft.com/en-us/power-bi/create-reports/deployment-pipelines-process#semantic-model-limitations

(Note that this link is the documentation listing unsupported scenarios under “Semantic model limitations.”)

How to resolve or work around it
Option 1 - Avoid the proxy model

Instead of connecting to another published Power BI semantic model:

Connect directly to the same data source in your model via DirectQuery or import

This removes the proxy layer and should let deployment pipelines work normally.

Option 2 - Import all tables

If the real-time DirectQuery connection to that other semantic model isn’t required:

Change those 4 DirectQuery tables to Import mode

Then your model isn’t composite and pipeline deployment will succeed.

Option 3 - Manual deployment

If you must keep your composite/proxy setup:

Publish the dataset manually to the target workspace (don’t use a pipeline)

This avoids the restrictions of deployment pipelines.

Not currently supported

There isn’t a setting you can tick to make pipelines support proxy models with DirectQuery/native queries it’s a platform limitation at this time.

If it solved your issue, feel free to mark it as the solution so others can benefit too.

Thanks for being part of the community.