Forum Discussion

kaspars_b's avatar
kaspars_b
Regular Visitor
1 year ago
Solved

Measuring Size on Disc for Direct Lake Semantic models

According Fabric capacity guardrails and limitations there are Max model size on disk/OneLake (GB) limitation for Direct Lake models for lower Fabric capacities. 

 

1) What is the most accurate way to measure this "Model size on disk" for an exisiting Direct Lake Semantic model?

2) Does Model size on disk depend on the number of columns currently transcoded into memory?

 

3) Or it primarily depends on Parquet file size in OneLake (Warehouse / Lakehouse) for tables used in Direct Lake model, even if none are Paged in currently?  

  • Hi kaspars_b 

     

    • The best way to measure the model size of a existing direct lake sematuc model is to check the szie of  parquet files stored in OneLake. for Direct Lake semantic models, Max Memory represents the upper memory resource limit for how much data can be paged in. 
    • Model sizee doesn't directly depend on number of columns currently transcoded into memory. size of the disck committed by the parquet files stored in onelake. If  the Max model size on disk/OneLake causes all queries to the semantic model to fall back to DirectQuery mode. All other guardrails presented in the table are evaluated per query.
    • Yes you are correct. the model size primarly depends on the size of parquest files stored in onelake.  Your Fabric SKU determines the maximum available memory per Direct Lake semantic model for the capacity. When the limit is exceeded, queries to the semantic model might be slower due to excessive paging in and out of the model data.

    Pls have a look on below snap to know about Fabric capacity guardrails 

     

    If you need more info pls go through below documentation. It might helps you

     

    Understand storage for Direct Lake semantic models - Microsoft Fabric | Microsoft Learn

    Direct Lake overview - Microsoft Fabric | Microsoft Learn

    Power BI usage scenarios: Prototyping and sharing - Power BI | Microsoft Learn

     

     

    Thank you!!!

     

    Did I answer your question? Mark my post as a solution!

    Proud to be a Super User!

     

  • Hi kaspars_b 

     

    yes, for F4 capacity total size of all Parquet files in Onelake for all tables used in  direct Lake semantic model can't exceed 10GB

     

    For workspace settings, sysytem storage it is irrelevent for measuring the model size on disck for sematic direct lake sematic models. System stoareg will reflects the overall storgae useage of workspace.

     

    Thank you!!!

     

    Did I answer your question? Mark my post as a solution!

    Proud to be a Super User!

6 Replies

  • v-ssriganesh's avatar
    v-ssriganesh
    Icon for Community Support rankCommunity Support

    Hi kaspars_b,

    Thank you for posting your query in the Microsoft Fabric Community Forum, and thanks to suparnababu8 for sharing valuable insights.

     

    Could you please confirm if your query has been resolved by the provided solution? If so, please mark it as the solution. This will help other community members solve similar problems faster.

    Thank you.

  • Hi kaspars_b 

     

    • The best way to measure the model size of a existing direct lake sematuc model is to check the szie of  parquet files stored in OneLake. for Direct Lake semantic models, Max Memory represents the upper memory resource limit for how much data can be paged in. 
    • Model sizee doesn't directly depend on number of columns currently transcoded into memory. size of the disck committed by the parquet files stored in onelake. If  the Max model size on disk/OneLake causes all queries to the semantic model to fall back to DirectQuery mode. All other guardrails presented in the table are evaluated per query.
    • Yes you are correct. the model size primarly depends on the size of parquest files stored in onelake.  Your Fabric SKU determines the maximum available memory per Direct Lake semantic model for the capacity. When the limit is exceeded, queries to the semantic model might be slower due to excessive paging in and out of the model data.

    Pls have a look on below snap to know about Fabric capacity guardrails 

     

    If you need more info pls go through below documentation. It might helps you

     

    Understand storage for Direct Lake semantic models - Microsoft Fabric | Microsoft Learn

    Direct Lake overview - Microsoft Fabric | Microsoft Learn

    Power BI usage scenarios: Prototyping and sharing - Power BI | Microsoft Learn

     

     

    Thank you!!!

     

    Did I answer your question? Mark my post as a solution!

    Proud to be a Super User!

     

    • kaspars_b's avatar
      kaspars_b
      Regular Visitor

      Thanks for a detailed response. 

      So for example for F4 capacity, the total size of all Parquet files in Onelake for all tables used in a Direct Lake semantic model cannot exceed 10GB. Correct? 

       

      And the Size that I see for this Direct Lake Semantic model when going to Workspace Settings ->System Storage is irrelevent here?

       

      • suparnababu8's avatar
        suparnababu8
        Icon for Super User rankSuper User

        Hi kaspars_b 

         

        yes, for F4 capacity total size of all Parquet files in Onelake for all tables used in  direct Lake semantic model can't exceed 10GB

         

        For workspace settings, sysytem storage it is irrelevent for measuring the model size on disck for sematic direct lake sematic models. System stoareg will reflects the overall storgae useage of workspace.

         

        Thank you!!!

         

        Did I answer your question? Mark my post as a solution!

        Proud to be a Super User!

  • v-ssriganesh's avatar
    v-ssriganesh
    Icon for Community Support rankCommunity Support

    Hi kaspars_b,

    May I ask if you have resolved this issue? If so, please mark the helpful reply and accept it as the solution. This will be helpful for other community members who have similar problems to solve it faster.

    Thank you.

  • v-ssriganesh's avatar
    v-ssriganesh
    Icon for Community Support rankCommunity Support

    Hi kaspars_b,
    I wanted to check if you had the opportunity to review the information provided. Please feel free to contact us if you have any further questions. If my response has addressed your query, please accept it as a solution and give a 'Kudos' so other members can easily find it.
    Thank you.