Performance Analyzer
Power BI · Advanced Features
Theory
Performance Analyzer in Power BI Desktop records how long each visual takes when a page is rendered or a user interacts. You open the pane, click Start recording, then refresh visuals or click slicers. For each visual it lists the total duration broken into parts: DAX query (time for the engine to return the data), Visual display (time to draw the visual) and Other (time waiting, for example behind other visuals or for background work).
The pattern of numbers tells you where to look. A long DAX query points to an expensive measure or model design, such as a FILTER over a whole fact table or many-to-many relationships. A long Visual display points to too many data points or a heavy custom visual. A long Other often means too many visuals on the page. You can copy the query of any visual and run it in DAX query view or an external tool such as DAX Studio to test improvements.
Why we use it
Slow reports frustrate users and get abandoned. Performance Analyzer replaces guesswork with measured timings, so you fix the actual bottleneck.
When to use it
Use it before publishing any important report and whenever users complain about speed. Test with a cold start using Refresh visuals, and compare before and after each change. For refresh (not rendering) slowness, look at Power Query and source performance instead.
How it works
- Report page feels slow
- Start recording and refresh visuals
- Find the visual with the longest duration
- Is the DAX query the largest part? Yes: Optimise the measure or model. No: Reduce visuals or data points.
- Record again and compare
- Page renders faster
Syntax
Desktop: Optimize ribbon → Performance analyzer (older versions: View → Performance analyzer)
Start recording → Refresh visuals / interact → Stop
Per visual: DAX query | Visual display | Other → Copy query / Run in DAX query viewExample
1. Open the Sales Overview page → Performance analyzer → Start recording → Refresh visuals.
2. The bar chart's DAX query time is much longer than the other visuals.
3. Copy query: the measure uses CALCULATE([Total Sales], FILTER(Sales, Sales[Region] = "North")).
4. Rewrite it as CALCULATE([Total Sales], Sales[Region] = "North") (column filter, not a table scan).
5. Record again and compare the DAX query time.Key points
- Records DAX query, Visual display and Other time per visual
- Long DAX query → optimise measures or the model
- Long Visual display → fewer data points or simpler visuals
- Long Other → too many visuals competing on the page
- Copy the query to test in DAX query view or DAX Studio