Timeout error when refreshing semantic model connected to dataverse

[shimgarah]

Hi everyone,

I'm experiencing a persistent issue with a semantic model in Power BI (both Desktop and Service) that connects to Dataverse using Import mode via Power Query.

Everything used to work fine — the refresh would complete in about 10 minutes. However, for the past few weeks, the refresh never finishes and eventually fails after around 40 minutes with the following error:

Microsoft SQL: The timeout period elapsed prior to completion of the operation or the server is not responding.

All my tables are connected to Dataverse, and the issue seems to occur during the loading of one or more specific tables. I’ve already tried:

• Reducing the number of records (I deleted about half of the rows).
• Removing unnecessary columns.
• Testing the refresh in Power BI Desktop (same result).
• Checking for complex relationships or transformations in Power Query.

Despite these efforts, the issue persists. I would really appreciate any ideas, suggestions, or guidance on what else I could try — especially if someone has faced a similar situation.

Thanks in advance! 

[rohit1991 · Accepted solution]

Hi @shimgarah ,

Since the issue started recently and used to work fine before, it may be tied to backend performance changes, query complexity, or API throttling on the Dataverse side.

>What Might Be Causing It

• API Throttling or Performance Issues on Dataverse side — particularly if your model is querying multiple large tables.

• Increased Query Complexity, such as joins, calculated columns, or dynamic filtering in Power Query.

• Network Latency or Capacity Limits from Power BI Service when working with Dataverse over large volumes of data.

• Power Query Transformations may still be folding inefficiently or not folding at all, especially with Dataverse’s API limitations.

>Steps You Can Try

1. Review Query Folding

   • In Power Query, right-click a query step and choose "View Native Query" — if this option is greyed out, your query isn’t folding, which can severely impact performance.

   • Try moving filters, column removals, and joins as early as possible in the query to encourage folding.

2. Split Large Tables

   • If one table is the bottleneck, try splitting it into two queries (e.g., current year + historical), and import them separately. You can merge or append them later if needed.

3. Use Incremental Refresh (If on Premium/Pro + Service)

   • Set up incremental refresh so that Power BI only loads changed data rather than reloading everything every time.

4. Monitor Dataverse Limits

   • Dataverse enforces API call limits and performance throttling.

   • Try monitoring request load through Power Platform Admin Center, or reduce the refresh frequency temporarily.

5. Use Dataflows (Optional)

   • Consider offloading the load process by using Power BI Dataflows to stage your Dataverse data, then connect your semantic model to that instead.

[shimgarah · Accepted solution]

Thank you all for your responses. The issue hasn't been resolved, but for now, I'm applying an incremental load that creates daily records in the table and updates those that have changed within the past year. However, if I want to modify any part of the design in Power BI Desktop, I will do so using only the initial data and will need to increase the duration of the initial load.

To implement the incremental load, I used the following tutorials:

• Tutorial 1: https://www.youtube.com/watch?v=yvUNDJo6uIs&t=1015s
• Tutorial 2: https://www.youtube.com/watch?v=0vbxtT4D--Y&t=28s
• Microsoft Learn resources: Actualización incremental de los modelos semánticos en Power BI - Power BI | Microsoft Learn

[v-dineshya]

Hi @shimgarah ,

Thank you for reaching out to the Microsoft Community Forum.

Hi @rohit1991 , @Poojara_D12 , @lbendlin  Thank you for your prompt response.

Hi @shimgarah , in addition to @rohit1991 , @Poojara_D12  response, I am adding some more points and Microsoft document for troubleshooting.

As you mentioned that, the refresh process which previously completed in about 10 minutes, now fails after around 40 minutes with the below error.

"Microsoft SQL: The timeout period elapsed prior to completion of the operation or the server is not responding".

This typically occurs during the loading of one or more specific tables from Dataverse.

Please refer below workarounds.

1. Update and scale out the on-premises data gateway.

2. Optimize the SQL queries for efficiency.

3. Remove unused columns/tables and avoid complex calculated columns.

4. Increase the SQL timeout in Power Query connection settings.

5. Clear the cache, restarting database service and gateway, and reload the problematic table it resolved the issue.

Please refer Microsoft official document.

Troubleshoot refresh scenarios - Power BI | Microsoft Learn

I hope this information helps. Please do let us know if you have any further queries.

Regards,

Dinesh

[v-dineshya]

Hi @shimgarah ,

We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet.And, if you have any further query do let us know.

Regards,

Dinesh

[rohit1991 · Accepted solution]

Hi @shimgarah ,

Since the issue started recently and used to work fine before, it may be tied to backend performance changes, query complexity, or API throttling on the Dataverse side.

>What Might Be Causing It

• API Throttling or Performance Issues on Dataverse side — particularly if your model is querying multiple large tables.

• Increased Query Complexity, such as joins, calculated columns, or dynamic filtering in Power Query.

• Network Latency or Capacity Limits from Power BI Service when working with Dataverse over large volumes of data.

• Power Query Transformations may still be folding inefficiently or not folding at all, especially with Dataverse’s API limitations.

>Steps You Can Try

1. Review Query Folding

   • In Power Query, right-click a query step and choose "View Native Query" — if this option is greyed out, your query isn’t folding, which can severely impact performance.

   • Try moving filters, column removals, and joins as early as possible in the query to encourage folding.

2. Split Large Tables

   • If one table is the bottleneck, try splitting it into two queries (e.g., current year + historical), and import them separately. You can merge or append them later if needed.

3. Use Incremental Refresh (If on Premium/Pro + Service)

   • Set up incremental refresh so that Power BI only loads changed data rather than reloading everything every time.

4. Monitor Dataverse Limits

   • Dataverse enforces API call limits and performance throttling.

   • Try monitoring request load through Power Platform Admin Center, or reduce the refresh frequency temporarily.

5. Use Dataflows (Optional)

   • Consider offloading the load process by using Power BI Dataflows to stage your Dataverse data, then connect your semantic model to that instead.

[shimgarah]

Thank you, I applied point 3

[Poojara_D12]

Hi @shimgarah 

You're facing a common yet frustrating issue that can arise when working with large or complex datasets from Dataverse in Import mode within Power BI. The error message — “Microsoft SQL: The timeout period elapsed prior to completion of the operation or the server is not responding” — typically points to a backend bottleneck, likely due to query folding breakdown, inefficient API calls, or Dataverse throttling. While your optimizations like reducing row counts and removing columns are valid steps, they might not fully address the root cause if the problem lies deeper in how Power Query communicates with Dataverse.

Dataverse uses OData feeds behind the scenes, and these can be particularly sensitive to filters, data volume, and server performance. If even one query fails to fold properly (i.e., it cannot be translated into a server-side query and instead retrieves all rows before applying transformations), performance can degrade dramatically. Use the "View Native Query" option in Power Query to check whether folding is occurring for each table, and ensure any filters or transformations are placed early and preserved before folding breaks. If folding is lost, try rewriting steps or splitting them into intermediate queries.

Another area to examine is Dataverse API limitations and throttling, especially if the data is large or if multiple refreshes are happening concurrently across environments. In such cases, breaking the refresh into smaller queries using parameters or even splitting the semantic model into smaller models can help.

If this issue persists in Power BI Service, consider using a dataflow as an intermediary layer, allowing the heavy data retrieval to be done separately from the report model. Also, make sure the gateway (if you're using one) and Power BI Service are using the latest versions, and check with your Dynamics/Dataverse admin for any recent changes in performance or throttling policies.

Ultimately, the combination of native query folding, data volume control, and service layer diagnostics (like Fiddler or Performance Analyzer in Power BI Desktop) can help isolate the exact step where performance drops or failures occur. If needed, raising a Microsoft support ticket with trace logs could also reveal underlying throttling or performance caps that aren’t visible through Power BI alone.

[lbendlin]

You have done your due diligence. Time to raise a ticket. Since you have a Pro license you can open a Pro ticket at https://admin.powerplatform.microsoft.com/newsupportticket/powerbi
Otherwise you can raise an issue at https://community.fabric.microsoft.com/t5/Issues/idb-p/Issues .