Forum Discussion
Mirrored Dataverse in Fabric
Hi,
in Fabric does it exist any native item to implement a Mirrored Dataverse to read/get data from this storage?
A such feature could be very useful.
Many thanks
Hi pmscorca,
Interesting question honestly, because the Dataverse + Fabric integration story is getting richer very quickly… but also a bit confusing sometimes with all the different concepts around, Fabric Link, Mirroring, OneLake exposure, Shortcuts, SQL endpoints Lakehouses.
As far as I know today, there is not yet a dedicated native Fabric item whose only purpose is something like "browse/query mirrored Dataverse storage directly" in the same way you might expect a standalone connector or explorer experience. That said, Fabric can absolutely consume Dataverse data natively today through different integration paths, depending on the architecture and setup.In practice, the two main paths are :
- Fabric Link (via Azure Synapse Link for Dataverse), the most stable and production-ready option today, exposes Dataverse tables as Delta tables in a Lakehouse.
- Dataverse Mirroring, currently in preview, more native Fabric experience but still evolving.
From there, Dataverse data is usually consumed through Lakehouse tables, SQL analytics endpoints, shortcuts, notebooks, or semantic models rather than through a separate dedicated "Dataverse mirrored reader" item.
So I think your idea actually makes a lot of sense
A more explicit/native experience for exploring and governing mirrored Dataverse data inside Fabric would probably make things much clearer for many users. Right now, the ecosystem is evolving pretty fast, and Microsoft seems to be progressively converging operational and analytical workloads into the same platform experience.Useful references:
- https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-view-in-fabric
- https://learn.microsoft.com/en-us/fabric/mirroring/overview
- https://learn.microsoft.com/en-us/fabric/onelake/onelake-shortcuts
- https://learn.microsoft.com/en-us/fabric/data-engineering/lakehouse-overview
- https://learn.microsoft.com/en-us/power-apps/maker/data-platform/fabric-link-faq
Hope this helps clarify the current landscape a little bit
And honestly, I would not be surprised if Microsoft expands the native Dataverse/Fabric experience further over time.
9 Replies
Hipmscorca,
asTamanchu mentioned in his nice answer, the topic is evolving fast. By now Microsoft has added to the mention link via Synapse also a direct link. Check out this docs for more info and also a compare to the other options.
https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-view-in-fabric
Best regards!
Hi pmscorca,
Interesting question honestly, because the Dataverse + Fabric integration story is getting richer very quickly… but also a bit confusing sometimes with all the different concepts around, Fabric Link, Mirroring, OneLake exposure, Shortcuts, SQL endpoints Lakehouses.
As far as I know today, there is not yet a dedicated native Fabric item whose only purpose is something like "browse/query mirrored Dataverse storage directly" in the same way you might expect a standalone connector or explorer experience. That said, Fabric can absolutely consume Dataverse data natively today through different integration paths, depending on the architecture and setup.In practice, the two main paths are :
- Fabric Link (via Azure Synapse Link for Dataverse), the most stable and production-ready option today, exposes Dataverse tables as Delta tables in a Lakehouse.
- Dataverse Mirroring, currently in preview, more native Fabric experience but still evolving.
From there, Dataverse data is usually consumed through Lakehouse tables, SQL analytics endpoints, shortcuts, notebooks, or semantic models rather than through a separate dedicated "Dataverse mirrored reader" item.
So I think your idea actually makes a lot of sense
A more explicit/native experience for exploring and governing mirrored Dataverse data inside Fabric would probably make things much clearer for many users. Right now, the ecosystem is evolving pretty fast, and Microsoft seems to be progressively converging operational and analytical workloads into the same platform experience.Useful references:
- https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-view-in-fabric
- https://learn.microsoft.com/en-us/fabric/mirroring/overview
- https://learn.microsoft.com/en-us/fabric/onelake/onelake-shortcuts
- https://learn.microsoft.com/en-us/fabric/data-engineering/lakehouse-overview
- https://learn.microsoft.com/en-us/power-apps/maker/data-platform/fabric-link-faq
Hope this helps clarify the current landscape a little bit
And honestly, I would not be surprised if Microsoft expands the native Dataverse/Fabric experience further over time.Hi, thanks for your reply.
Link to Fabric is a feature of Dataverse, not Fabric. It creates separate "mirrored tables" inside Dataverse, but the Dataverse storage has a significant cost.
So, in terms of costs, using Link to Fabric feature is not the best solution.
ThanksHi pmscorca,
Yes, you are correct.
Based on the current Microsoft documentation, the Fabric integration for Dataverse still relies on the Dataverse Managed Lake / Link to Fabric architecture.
The official documentation states:
“Dataverse tables must first be available in the Dataverse Managed Lake before they're visible in the Fabric create shortcuts UX.”
So currently, even when creating Dataverse shortcuts from Fabric, the tables must already exist in the Dataverse Managed Lake first.
Because of this architecture, your concern regarding Dataverse storage usage with Link to Fabric is understandable, especially for large environments.
For more details, Please go through offical documentation:
https://learn.microsoft.com/en-us/fabric/onelake/create-dataverse-shortcut#create-shortcuts-through-fabric
I hope this helps. Please feel free to reach out if you have any further questions.
Thank you.
Hi,
in order to recap:
- using Link to Fabric feature is expensive due to Dataverse storage cost;
- Dataverse Mirroring in Fabric is in working (when could it be released?);
- alternative good solution?
Thanks
Hi pmscorca,
As an alternative approach, you can consider using Dataflows Gen2, Pipelines, notebooks to move only the required Dataverse tables/data into OneLake.
This can provide more control over the amount of data being ingested and stored.
For roadmap or future feature visibility, you can also monitor:
https://roadmap.fabric.microsoft.com/And if a more storage-efficient native Dataverse integration in Fabric is important for your scenario, submitting an idea in the Ideas Forum may help increase visibility for the request:
https://ideas.fabric.microsoft.com/Thank you.
