Forum Discussion
Direct Lake Model Based Report Spikes Crazy CU's
I created a Direct Lake Sementic Model that is build on top of Fabric Warehouse. in which, Fact table Contains close to 800 Million rows. also I built Dashboard using that Sementic model. when anyone view that dashboard it spikes report capacity over 100% just one report by one user view.
Tried doing performance analyzer and fix long running DAX that was taking long time. and now all the visuals max it takes is 2 seconds. So I am not able to find out where this CU spike comes from. I used Capacity metrics app and see time in point detail report it just says query. but not able to see what exact query is.
Can anyone help me what is the right way to trouble shoot interactive CU's spike on report from Direct Lake sementic model. So it can be performant so we will have no resources issue. BTW
What you could do is open up your Power BI report in Power BI Desktop and then use the performance analyzer on your page to see which query is taking the longest to return. That potentially could be the query that is causing the spark in your capacity also. If you don't have any filters in place and you have something like a table that has to query all 800 million rows, it has to load those 800 million rows. Into memory, which can cause a spike.
Hi RaviPatel0007 ,
Thank you for reaching out to Microsoft Fabric Community.
Thank you GilbertQ for the prompt response.
Since visuals now load quickly, the CU spike is likely caused by DirectQuery fallback - Power BI may be switching from Direct Lake to DirectQuery behind the scenes, which uses more capacity.
You can try to check:Use DAX Studio or Query Diagnostics to confirm if fallback is happening.
Consider setting DirectLakeOnly to force Direct Lake mode.
Simplify your model - reduce columns, avoid views, and use star schema design.
The Capacity Metrics App shows “query” but not exact DAX - use Performance Analyzer in Desktop to match visuals to queries.Thank you.
5 Replies
What you could do is open up your Power BI report in Power BI Desktop and then use the performance analyzer on your page to see which query is taking the longest to return. That potentially could be the query that is causing the spark in your capacity also. If you don't have any filters in place and you have something like a table that has to query all 800 million rows, it has to load those 800 million rows. Into memory, which can cause a spike.
Hi RaviPatel0007 ,
Thank you for reaching out to Microsoft Fabric Community.
Thank you GilbertQ for the prompt response.
Since visuals now load quickly, the CU spike is likely caused by DirectQuery fallback - Power BI may be switching from Direct Lake to DirectQuery behind the scenes, which uses more capacity.
You can try to check:Use DAX Studio or Query Diagnostics to confirm if fallback is happening.
Consider setting DirectLakeOnly to force Direct Lake mode.
Simplify your model - reduce columns, avoid views, and use star schema design.
The Capacity Metrics App shows “query” but not exact DAX - use Performance Analyzer in Desktop to match visuals to queries.Thank you.
Hi RaviPatel0007 ,
I wanted to check if you had the opportunity to review the information provided and resolve the issue..?Please let us know if you need any further assistance.We are happy to help.
Thank you.
Hi RaviPatel0007 ,
May I ask if you have resolved this issue? Please let us know if you have any further issues, we are happy to help.
Thank you.
Hi RaviPatel0007 ,
I hope the information provided is helpful.I wanted to check whether you were able to resolve the issue with the provided solutions.Please let us know if you need any further assistance.
Thank you.