Memory issue
Hello community I get this error on a table visual in the Power BI Service in a Table visual : The context is the following Table Source : ~600M rows, Import mode, loaded from Snowflake. Capacity: F64. The visual only uses columns from the table (no DAX, no calculated columns). The query is over the limit as you can see. Is there a way to stay under it while keeping a single visual? Can the per-query memory limit be raised on F64, or is 10 GB the maximum for this SKU? Thank you!
Power BI Service Returning Different Results Than Desktop After Refresh
a { text-decoration: none; color: #464feb; } tr th, tr td { border: 1px solid #e6e6e6; } tr th { background-color: #f5f5f5; } Hi all, I have a Power BI report where the calculations and visuals are fully validated in Power BI Desktop, but after publishing and refreshing the semantic model in Power BI Service, the results change despite the underlying source data appearing to be identical. What I've verified The report calculates correctly in Desktop. The semantic model is refreshed successfully in Power BI Service. Source totals match between Desktop and Service. Relationships are active and functioning. Diagnostic measures return the expected values in Desktop. The same diagnostic measures can also return the expected values in Service. The issue A measure that calculates a monthly ratio (Cost ÷ Volume) returns the expected values in Desktop, but after refresh in Power BI Service the same visual displays different monthly values. What's particularly confusing is: The overall totals reconcile correctly. The underlying source totals reconcile correctly. The discrepancy only appears in the monthly breakdown. Service returns the same incorrect monthly values consistently after refresh. Republishing a freshly refreshed PBIX restores the expected values. A subsequent semantic model refresh in Service causes the monthly values to change again. Additional observations We tested diagnostic measures separately and confirmed: Numerator values are correct. Denominator values are correct. Benchmark months reconcile correctly in Desktop. The semantic model contains the correct overall totals. The issue appears only after Service refresh and only affects the monthly allocation of values. Question Has anyone encountered a scenario where: Desktop calculations are correct, Service refresh succeeds, overall totals remain identical, but monthly allocations differ between Desktop and Service after refresh? Are there any known issues related to: Service refresh evaluation differing from Desktop, date grouping or hierarchy evaluation, semantic model refresh behaviour, stale visual metadata, report-to-semantic-model binding, or Excel/SharePoint sourced data being interpreted differently during Service refresh? Any troubleshooting suggestions would be greatly appreciated as we've verified the calculations extensively in Desktop and the issue appears isolated to how the refreshed model behaves in Power BI Service.
Premium P3 – Same Interactive Operation Consumes 0.25% vs. 25% of Base Capacity During Overload
Hello Power BI Community, I'm investigating an unusual behaviour regarding interactive capacity consumption and report performance in Power BI Premium, and I would appreciate your insights. My Report is: Capacity: Power BI Premium P3 Semantic Model: Import mode Report: Published in Power BI Service Monitoring: Microsoft Fabric Capacity Metrics App Under normal capacity conditions, the report performs exceptionally well, with most visual interactions completing within 1–2 seconds. However, when the Premium capacity becomes overloaded, the same report interactions can take up to 5 minutes to complete. While the significant increase in response time could potentially be explained by capacity throttling, queuing, or resource contention, what I find particularly unusual is the reported capacity consumption. In the Fabric Capacity Metrics App, specifically looking at "% of Base Capacity" for interactive operations, I observe the following: Metric Normal Capacity Overloaded Capacity Report interaction response time 1–2 seconds Up to 5 minutes % of Base Capacity (Interactive) ~0.25% ~25% Semantic model storage mode Import Import User interaction Same Same In other words, the same report interaction appears to account for approximately 100 times more of the base capacity during overload. I need your help to understand the following: How can the same interactive operation account for such a significant difference in "% of Base Capacity"? Is this metric influenced by capacity utilisation, smoothing, throttling, or the time window over which consumption is calculated? Does the "% of Base Capacity" metric represent actual CU consumption attributable to the operation, or a normalised percentage influenced by other capacity activities? Could resource contention, query retries, memory pressure, or changes in execution behaviour during overload result in genuinely higher CU consumption for identical DAX queries? Is it possible that the increased percentage is related to the attribution or smoothing of interactive CU consumption rather than an actual 100-fold increase in computational work? What would be the recommended approach to distinguish increased query execution cost from capacity-induced waiting time? Are there specific metrics, detailed operation views, or diagnostic tools that could help identify the root cause? How can I help with capacity consumption? What should I review/change? Please take into consideration that the report is already optimised, and when the capacity is not overloaded, it performs smoothly with minimum consumption. It is the report with the greatest number of unique users in the Capacity (591). My main concern is not only the performance degradation during overload, but whether the same report operation actually becomes significantly more resource-intensive, or whether the Capacity Metrics App presents consumption differently under these conditions. Any technical explanation, documentation, or recommendations on how to investigate this behaviour would be greatly appreciated. Thank you in advance for your help! EleftheriaAppending Oracle (Gateway) and Dataverse (Service Principal) fails refresh in Service
Hi everyone, I'm facing a persistent refresh issue in Power BI Service that works perfectly in Desktop but fails after publishing. I need help understanding the correct architecture or workaround. My Setup: Source A (Oracle): On-premises database. Connected via an On-premises Data Gateway. Authentication is Basic (username/password). Source B (Dataverse): Cloud source. Connected via Service Principal (or Organizational account). Transformation: In Power Query (Desktop), I need to Append (Union) tables from Oracle and Dataverse into a single table for analysis. There are no complex calculations, just row-level data for reporting. The Problem: When I publish the semantic model to Power BI Service and attempt a refresh, it fails. The error message indicates a Mashup/Data Access issue. It appears that Power BI Service tries to route the Dataverse cloud connection through the On-premises Gateway to combine it with Oracle. Even after enabling the gateway setting "Allow user's cloud data sources to refresh through this gateway cluster," the refresh still fails, or Dataverse incorrectly shows up under the Gateway binding in the semantic model settings. I have tried clearing permissions in Desktop and republishing, but the conflict persists. Removing the Append step and keeping tables separate works, but I specifically need the union of these rows. My Question: What is the correct, supported way to append (union) an on-premises Oracle table and a cloud Dataverse table in a single semantic model that refreshes successfully in Power BI Service? Is Dataflow Gen2 the only viable solution (combining the sources there and then consuming the Dataflow in the model)? If so, are there specific licensing or permission requirements for the Service Principal to access Dataverse within a Dataflow? Are there any known limitations with mixing Service Principal authentication (for Dataverse) and Gateway authentication (for Oracle) within the same Power Query mashup? Any guidance or workarounds would be greatly appreciated. Thank you!
Power BI Dataflow Refresh Fails Despite Queries Loading Successfully in Power Query
Hi everyone, I'm facing an issue with a Power BI Dataflow and would appreciate some guidance. I have three queries pulling data from different sources, which I've appended into one final query using Table.Combine. All three queries load correctly in Power Query, and the appended query also previews the combined data without any errors. I've disabled Enable Load for the three source queries and kept it enabled only for the final appended query. However, whenever I refresh the dataflow, the refresh fails, and the error only shows a Request ID and Activity ID without any specific details. I've also tried simplifying the final query to just the basic append operation, but the refresh still fails. Has anyone experienced something similar? Could this be related to query dependencies, combining different data sources, or dataflow settings? Any suggestions on how to troubleshoot or resolve this would be greatly appreciated. Thank you!
Translytical task flow gets 401 on generatemwctokenv2 when embedded in Salesforce, Entra SSO
We have a Power BI report with a translytical task flow button that calls a Fabric user data function to write to a Fabric warehouse. It works in the Power BI service. We embed the same report in a Salesforce Lightning page with the Power BI JavaScript SDK, and the button fails there. What we see App-owns-data (service principal embed token). Clicking the button sends POST /metadata/v201606/generatemwctokenv2 with Authorization: EmbedToken ... and x-powerbi-hostenv: Embed for Customers. The response is 401 with an empty body, and the data function is never called. User-owns-data (Entra SSO, our own app registration). The same call now carries Authorization: Bearer ... and x-powerbi-hostenv: Embed for Organization, and still returns 401. The decoded token has the Power BI API audience, idtyp: user, MFA, scp: Report.Read.All, and the appid of our own app registration. Power BI service (works). The same call carries a Bearer token with scp: user_impersonation and a different appid, which looks like Microsoft’s own Power BI app. The call succeeds. So the identity type changed between 1 and 2 but the 401 remained. The scopes and the issuing app are the remaining differences from the working case. What we’re testing next We’re adding UserDataFunction.Execute.All and Fabric.Embed as delegated permissions on our app registration, granting admin consent, and retesting with a fresh token. Questions Does generatemwctokenv2 accept access tokens issued to a customer-registered Entra app, or only tokens from Microsoft’s own Power BI app? If custom app tokens are supported, which delegated scopes are required for a task flow to run? Are task flows supported at all with app-owns-data embed tokens? Has anyone run a task flow button inside an embedded report outside the Power BI service, and in which embedding mode? Kindly let me know in case of any additional information needed.
"We could not evaluate this query due to invalid or missing credentials" with Business Central
Hello everyone, I recently started receiving the following error when refreshing a Power BI Desktop report connected to Dynamics 365 Business Central Online: "We could not evaluate this query due to invalid or missing credentials." Environment: Power BI Desktop (latest version) Data source: Dynamics 365 Business Central Online Using the standard Dynamics 365 Business Central connector Authentication: Organizational Account No custom APIs or OData feeds Query contains only the Source step Has anyone recently experienced the same issue with the Dynamics 365 Business Central connector? I have also noticed reports from other users mentioning intermittent credential failures with the Business Central connector, and it appears Microsoft may already be investigating a broader issue. Any guidance or confirmation would be appreciated. Thanks!Mirrored SharePoint Online List only shows system/default lists, not my custom SharePoint lists
Hi everyone I'm new to using Power BI and I've run in to a problem that I can't find an answer to. I'm trying to create a Mirrored SharePoint Online List in Power BI. The connection to the SharePoint site succeeds, but when I reach "Choose data", my actual SharePoint lists are not displayed. Instead, I only see generic/system objects such as: CompanyList-... CSPViolationReportList Documents Form Templates SharePointHomeOrgLinks TaxonomyHiddenList Typography Library Site Assets This list does not match what I see on SharePoint. I have tested this with different SharePoint sites, but I keep seeing the same type of generic/system lists rather than the custom lists contained in the selected SharePoint site.The SharePoint lists themselves are accessible to my account. I can't figure out why i cannot access my list from SharePoint. Any help and suggestions are greatly appriciated :) Thank you
