Forum Discussion

Luigia-Costabil's avatar
Luigia-Costabil
Advocate I
2 months ago
Solved

Fabric Dataverse - Synapse Link versus Fabric Link

For a new Dataverse + Microsoft Fabric project, which approach would you recommend today and why?

  • Link to Microsoft Fabric

  • Azure Synapse Link + ADLS + Fabric Shortcut

My main concerns are:

  • Cost

  • SCD Type 2 / CDC support

  • Schema evolution

  • Long-term maintainability

I'm interested in real-world experiences and lessons learned from production implementations. Thank you!

  • Tamanchu's avatar
    Tamanchu
    2 months ago

    Good follow-up question.

    For the pattern you described :

    Dataverse → Azure Synapse Link → ADLS Gen2 → Fabric Shortcut

    you don't necessarily need to deploy a full Azure Synapse Workspace if the goal is only to land the Dataverse data into Azure Data Lake Storage Gen2 and then consume it from Fabric through shortcuts.

    Microsoft documents an option to create Azure Synapse Link for Dataverse with Azure Data Lake Storage Gen2. In that setup, the key Azure component is the ADLS Gen2 storage account. A Synapse Workspace is a separate option, mainly needed if you also want to use Synapse Analytics capabilities directly.

    So the lighter architecture would be :

    Dataverse → Azure Synapse Link service → ADLS Gen2 → Fabric Shortcut

    A few things to keep in mindv:

    • The ADLS Gen2 account must be in the same Microsoft Entra tenant as the Power Apps tenant.
    • Hierarchical namespace must be enabled.
    • Required storage roles must be assigned correctly.
    • Change tracking must be enabled on the Dataverse tables you want to export.
    • If you need Delta/Parquet output, check the Delta Lake export requirements as well.

    Useful Ms docs:

    Azure Synapse Link for Dataverse with Azure Data Lake Storage Gen2 :
    https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-data-lake 

    Export Dataverse data in Delta Lake format :
    https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-delta-lake 

    Azure Synapse Link for Dataverse with Azure Synapse Analytics workspace :
    https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-synapse 

     

    So in short, if Fabric is already the analytics layer and you only need ADLS Gen2 as the landing zone, I would start with the ADLS Gen2-only option to keep the architecture lighter. A Synapse Workspace can always be added later if there is a real Synapse-specific requirement.

3 Replies

  • Hi Luigia-Costabil,

    This follows nicely from your previous question about Append Only and Link to Fabric.

    If I were starting a new Dataverse + Fabric project today, my recommendation would depend on the business requirements rather than choosing one approach for every scenario.

    • If the goal is analytics on the latest state of the data, with a simpler architecture, lower operational overhead, and staying aligned with the Fabric roadmap, I would choose Link to Microsoft Fabric.
    • However, if the project requires a true Bronze layer that preserves historical changes, delete events, or supports SCD Type 2 / CDC patterns, I would still lean towards Azure Synapse Link (Append Only) + ADLS + Fabric Shortcut. Based on the current Microsoft documentation, this remains the most clearly documented approach for preserving change history through the "IsDelete" column.

    So my decision matrix would be :

    • Latest-state analytics → Link to Microsoft Fabric
    • CDC / SCD Type 2 / historical replay → Azure Synapse Link (Append Only)

    My expectation is that Link to Fabric will continue to evolve, but today I haven't found official documentation stating that it provides an equivalent to the Append Only mode or the "IsDelete" column available with Azure Synapse Link.

     

    Useful Microsoft documentation:

    Link to Microsoft Fabric :
    https://learn.microsoft.com/power-apps/maker/data-platform/azure-synapse-link-view-in-fabric 

    Azure Synapse Link – In-place updates vs. Append Only writes :
    https://learn.microsoft.com/power-apps/maker/data-platform/azure-synapse-link-advanced-configuration#in-place-updates-vs-append-only-writes 

     

    I'd be interested to hear how others are approaching production implementations, especially for CDC and SCD Type 2 workloads.

    • Luigia-Costabil's avatar
      Luigia-Costabil
      Advocate I

      Thanks, that's very helpful.

      One follow-up question regarding the Azure Synapse Link option:

      In my scenario, the customer already has Microsoft Fabric, but there is currently no Azure Synapse environment deployed.

      If I choose the architecture:

      Dataverse → Azure Synapse Link → ADLS Gen2 → Fabric Shortcut

      would it be sufficient to provision only an Azure Data Lake Storage Gen2 account, or is an Azure Synapse Workspace also required?

      The reason I'm asking is that introducing a Storage Account is usually acceptable, but deploying and managing a dedicated Synapse Workspace solely for Dataverse integration would significantly increase the complexity and cost of the solution.

       

      Thanks!

      • Tamanchu's avatar
        Tamanchu
        Super User

        Good follow-up question.

        For the pattern you described :

        Dataverse → Azure Synapse Link → ADLS Gen2 → Fabric Shortcut

        you don't necessarily need to deploy a full Azure Synapse Workspace if the goal is only to land the Dataverse data into Azure Data Lake Storage Gen2 and then consume it from Fabric through shortcuts.

        Microsoft documents an option to create Azure Synapse Link for Dataverse with Azure Data Lake Storage Gen2. In that setup, the key Azure component is the ADLS Gen2 storage account. A Synapse Workspace is a separate option, mainly needed if you also want to use Synapse Analytics capabilities directly.

        So the lighter architecture would be :

        Dataverse → Azure Synapse Link service → ADLS Gen2 → Fabric Shortcut

        A few things to keep in mindv:

        • The ADLS Gen2 account must be in the same Microsoft Entra tenant as the Power Apps tenant.
        • Hierarchical namespace must be enabled.
        • Required storage roles must be assigned correctly.
        • Change tracking must be enabled on the Dataverse tables you want to export.
        • If you need Delta/Parquet output, check the Delta Lake export requirements as well.

        Useful Ms docs:

        Azure Synapse Link for Dataverse with Azure Data Lake Storage Gen2 :
        https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-data-lake 

        Export Dataverse data in Delta Lake format :
        https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-delta-lake 

        Azure Synapse Link for Dataverse with Azure Synapse Analytics workspace :
        https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-synapse 

         

        So in short, if Fabric is already the analytics layer and you only need ADLS Gen2 as the landing zone, I would start with the ADLS Gen2-only option to keep the architecture lighter. A Synapse Workspace can always be added later if there is a real Synapse-specific requirement.