Dashboard Refresh
The BI Pixie Dashboard is a Power BI template app that imports telemetry data from your BI Pixie storage into a semantic model. Each time the semantic model refreshes, it reads event data from your storage container and updates the dashboard with the latest insights.
This page explains how dashboard refresh works, what limits apply to each plan, and how to get the most out of your refresh budget.
How It Works
BI Pixie continuously collects telemetry from your tracked Power BI reports and stores it in your dedicated storage container. The dashboard is a separate Power BI app that you schedule to refresh on a regular cadence, just like any other Power BI semantic model.
Each refresh pulls event data from your storage container and loads it into the semantic model. The amount of data loaded depends on two factors:
- Data retention determines how long telemetry data is kept in your storage container.
- Dashboard refresh window (Pro and Enterprise) determines how many days of historical data the semantic model imports on each refresh.
For example, a Pro customer with 365-day retention and a 90-day refresh window will keep a full year of raw data in storage, but each dashboard refresh only loads the most recent 90 days into the semantic model.
Refresh Limits by Plan
| Trial | Standard | Pro | Enterprise | |
|---|---|---|---|---|
| Refreshes per day | 2 | 2 | 4 | 4 |
| Data retention (max) | 14 days | 30 days | 365 days | 1,095 days |
| Dashboard refresh window | All data | All data | 7 to 365 days | 7 to 365 days |
| Refresh strategy | Full scan | Full scan | Full scan | Full scan. For incremental refresh and Direct Lake over the full retention, move the pipeline to the Workload or Databricks (see below). |
The refresh window cannot exceed your current retention period. For example, if your retention is set to 30 days, the maximum refresh window is 30 days.
What Happens When You Reach the Limit
BI Pixie tracks how many times your dashboard semantic model is refreshed each day. If your dashboard is refreshed beyond the daily limit, further refreshes will be temporarily paused. You will see a clear message in Power BI explaining that the daily refresh limit has been reached.
No telemetry data is lost during this time. BI Pixie continues collecting events from your tracked reports as usual. Once the limit resets, your next scheduled refresh will pick up all the data that was collected in the meantime.
Tip: If you schedule your dashboard to refresh in Power BI Service, make sure the number of scheduled refreshes does not exceed your daily limit. For example, on a Standard plan with 2 refreshes per day, schedule the semantic model to refresh twice (e.g., morning and evening).
Optimizing Your Refresh
Use the Dashboard Refresh Window
On Pro and Enterprise plans, you can configure the dashboard refresh window in the Data Management page. This controls how many days of historical data the semantic model imports on each refresh.
A narrower refresh window means each refresh processes less data and completes faster. This is especially useful if you have a long retention period but only need the dashboard to show recent activity.
For example, a Pro customer tracking 10 reports with 365-day retention might set the refresh window to 90 days. The full year of raw data remains available for searches in the Data Management page, while the dashboard stays focused on the most recent quarter.
Schedule Refreshes Thoughtfully
You do not need to use all your daily refreshes. If your organization reviews the dashboard once a day, a single morning refresh may be sufficient. Save additional refreshes for days when you need more up-to-date data (e.g., after a major report launch).
Upgrade Your Plan
If you consistently need more refreshes or a longer refresh window, consider upgrading to a higher plan. See Plans and Billing for details.
Scaling beyond the import refresh
The import refresh above works well for most volumes, but if you retain years of telemetry or track a large number of reports, the fastest and most scalable approach is to move the telemetry pipeline to a Spark ETL that reads directly from your storage. Instead of importing the full history on every refresh, the ETL processes only new events and the dashboard reads the result through Direct Lake — with no import step and no refresh timeouts, regardless of how much data you keep. Two options deliver this:
- BI Pixie Workload for Microsoft Fabric. If you use Microsoft Fabric, the Workload deploys the ETL notebook and a Direct Lake dashboard into your own Fabric workspace automatically. This is the recommended path when you have a Fabric capacity.
- Power BI over Azure Databricks. If you already run Azure Databricks and do not use Fabric, the same ETL runs on your Databricks workspace and feeds the dashboard as an Import, DirectQuery, or Direct Lake model.
Both keep your telemetry in your own environment with full control over access and governance, and the pipeline can be extended with custom measures, reports, or AI agents. To find the right option for your setup, contact support.
Data Is Never Lost
Regardless of your plan or refresh configuration, BI Pixie never stops collecting telemetry from your tracked reports. Dashboard refresh limits only affect how often the Power BI semantic model is updated. Your raw telemetry data is always available in your storage container and can be searched from the Data Management page.