This time we’re going bigger than ever. Fabric, Power BI, SQL, AI and more. We're covering it all. You won't want to miss it.
Learn moreDid you hear? There's a new SQL AI Developer certification (DP-800). Start preparing now and be one of the first to get certified. Register now
Hello!
I need some assistance understanding something in the Capacity Metrics report. We have had a few spikes over 100%, and when I look at the timepoint detail for those, this is an example of what I see:
Just outside the screenshot, I see that for this one item, Query at this timepoint is consuming a siginificant percentage of the base capacity. I want to work with the report developer to assist them in optimizing their report to eliminate the cause of this overage.
What I need help with is how to direct them on identifying possible causes. Based on what I understood so far, "Query" means "Queries initiated by an on-demand data request from a user. For example, loading a model when opening a report, or user interaction with a report." I opened the report myself to test performance, and the report loads quickly, interactions with the visuals are quick and responsive. So it appears to me that loading a model when opening the report or user interaction with the report are not consuimg much CU.
What could I be missing here?
Thank you!
Based on user feedback (visuals rendering slowly) the developer needs to identify measures that perform poorly. These measures then need to be refactored. Goal should be a visual rendering time under 5 seconds.
There has been no feedback from users on slowly rendering visuals, and I just tried it myself again--the report fully rendered in 3 seconds.
This suggests to me one of three possibilities:
Does the User column say "Power BI Service" or does it list an actual user?
It's an actual user.
And that user did not notice that one of the visuals took over two minutes to render?
They didn't complain about it, but they could be one of those users who suffer in silence. 😀
I am planning to reach out to the user to ask about their experience, but I first wanted to make sure I was on the right track and that the Capacity Metrics were telling me what I thought they were telling me. Considering what happened for me, I was not at all confident that it meant what it looked like it meant.
Oh I could tell stories - about the astoundingly stupid mindboggling 100K limit, or about the Direct Queries marked as Background etc ...
Good luck on your journey.
Check out the April 2026 Power BI update to learn about new features.
Sign up to receive a private message when registration opens and key events begin.
If you have recently started exploring Fabric, we'd love to hear how it's going. Your feedback can help with product improvements.
| User | Count |
|---|---|
| 10 | |
| 8 | |
| 8 | |
| 8 | |
| 7 |
| User | Count |
|---|---|
| 49 | |
| 27 | |
| 24 | |
| 20 | |
| 20 |