Refresh error

[RTNL]

When refreshing the dataflow and semantic model I get the error shown below.

What could be the root cause of this issue?

Requested on,Dataflow name,Dataflow refresh status,Table name,Partition name,Refresh status,Start time,End time,Duration,Rows processed,Bytes processed (KB),Max commit (KB),Processor Time,Wait time,Compute engine,Error,

2026-01-08 21:41:14,RT Parts,Failed,Data,NA,Failed,2026-01-08 21:41:14,2026-01-08 21:43:45,00:02:31.3370,NA,NA,NA,NA,NA,NA,Error:  Request ID: 51a65b81-7a1b-b14b-c379-eff3f8333f1e Activity ID: 55feb5f9-6def-479c-81d0-d3e5f1581ec9

[v-dineshya]

HI @RTNL ,

Thank you for reaching out to the Microsoft Community Forum.

Hi @Poojara_D12 , Thank you for your prompt response.

HI @RTNL , In addition to @Poojara_D12  response, i have added some points.

1. Update to the latest gateway, ensure it’s online, and check all data sources are mapped correctly in the workspace.

2. Open the dataflow’s queries and check recent steps, Look for type conversions like Decimal --> Whole Number recently added. Check for text in numeric columns or locale/format mismatches like dates, decimals. If Excel/CSV is a source, remove trailing formulas that return #N/A in “blank” rows.

3. Turn Enhanced Compute OFF --> refresh --> then ON --> refresh again. This forces cache rebuild and can clear “NA/Cached” anomalies.

4. If the dataflow fails, the downstream dataset refresh always fails. Ensure the dataflow completes first.

Please refer below links.

Troubleshoot refresh scenarios - Power BI | Microsoft Learn

Solved: Dataflow refresh error wit Request ID and Activity... - Microsoft Fabric Community

Solved: DATAFLOW REFRESH COMPUTE ENGINE ERROR - Microsoft Fabric Community

Datasets - Get Refresh History - REST API (Power BI Power BI REST APIs) | Microsoft Learn

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

Regards,

Dinesh

[Poojara_D12]

Hi @RTNL 

This pattern (dataflow + semantic model both failing with only a Request ID and no table or partition listed) almost always means the failure happened before any table was actually processed, which points to a connection or authentication break rather than a data or DAX issue. In practice this is usually caused by expired or changed credentials, a revoked service principal, a rotated database password, or a gateway / VNet / firewall change that blocks the dataflow’s connection to the source. Because the engine cannot even open the source, it never gets far enough to identify a table, so everything shows as NA. Re-authenticate the dataflow connection and check gateway or network access — that almost always resolves this exact failure pattern.