Limitations and Refresh Timeouts

The BI Pixie Dashboard is implemented as a Power BI app, where its semantic model imports all BI Pixie usage data from your BI Pixie storage account on Azure. The implementation is based on Power Query and may experience timeouts during the refresh of the semantic model, when your organization generates a high volume of usage data across Power BI reports and dashboards.

Note: when timeout occurs during the refresh of BI Pixie Dashboard, you don't lose any usage data that BI Pixie collects. You can process BI Pixie data on your own and analyze it as you see fit, or contact us and we can guide you how to optimize BI Pixie and keep using our dashboard.

When the workspace you use for BI Pixie Dashboard is assigned to a Power BI Pro license, the semantic model's refresh time is limited to two hours. This is a Power BI platform limitation. If you encounter refresh timeouts due to large data volumes, you can follow one of the options below to resolve the issue.

We recommend that you open a support ticket, and our team will guide you through the best resolution for your environment.

Options to Resolve Refresh Timeout Issues

1. Limit the Data Range
   Change the Last N Days parameter of the semantic model from -1 to 30 to import only the last 30 days of data.
   Try refreshing the dashboard again to confirm whether the issue is resolved.

2. Heatmap Activation
   Did you activate the heatmap (as described here)? Consider turning it on only for specific reports, where this functionality is needed.

3. Upgrade to BI Pixie on Azure
   If you currently use BI Pixie on the Power Platform, consider upgrading to the Azure deployment.
   On Azure, the BI Pixie Dashboard imports data more efficiently through the Azure Storage Account. The Power Platform version was implemented for low usage in reports and is not efficient for large volume of usage due to its Dataverse connector.

4. Optimize Performance Parameters
   If you are already using the Azure deployment and have enabled the Performance feature, you can reduce refresh time by lowering the Last N Days Log Analytics parameter.
   The default is 30 days. Try using a smaller range and verify if it resolves the timeout.

5. Use the PBIX Semantic Model
   Customers on the Professional or Enterprise tiers who have access to the BI Pixie source code can use the provided PBIX file to refresh the data outside of the Power BI Pro service's one-hour limitation.
   However, for a more robust and scalable solution, see the next option.

6. Leverage Microsoft Fabric Capacity
   If you have a Fabric capacity, you can assign the BI Pixie workspace to a Fabric capacity to support larger data volumes and allow refresh operations with up to 24-hour timeouts. You can also implement incremental refresh for selected tables or deploy multiple BI Pixie dashboards filtered by project or workspace to distribute the load.

7. Deploy the Fabric Lakehouse Version (Enterprise)
   Enterprise customers can deploy the BI Pixie data pipeline as a Fabric Lakehouse with 20 Dataflows Gen2 in their own Fabric workspace. This architecture uses incremental refresh (only new telemetry is loaded on each run) and Direct Lake connectivity (the dashboard reads directly from the Lakehouse with no import step). This eliminates refresh timeouts entirely, regardless of data volume or retention period.
   
   See How We Deploy BI Pixie Dashboard Across Customers Using Fabric Dataflows for details on the architecture. Contact support to discuss enabling this for your environment.

8. Request our Professional Services to build your own optimization
   If options 1-7 are not sufficient to meet your requirements (e.g. you must process 3 years of data in less than 24 hours and want to use Spark compute to process the data instead of Dataflows Gen2), please open a support ticket and describe your requirements. We can offer you a paid training or professional services to help you build your analytics solution and process the collected data from BI Pixie.

The following table summarizes your options and support coverage: