Premium Capapcity Utilitization and Metrics performance delta high

Question

Hello,

In the Premium Capapcity Utilitization and Metrics I received high performance delta's on several reports, ranging from -96 to -42. All of the reports have been published to PowerBI service and the data source is from the users local machine. All of the reports are small.

I wanted to know if when the reports are run if all of the data goes to memory because it's on their local machine.

Is this setup the cause of the high delta's? Does this setup degrade performance when report data source is on local desktop and published to powerbi service?

Kind Regards,

v

Answer 1

Hi vw,

Thank you for reaching out to us on Microsoft Q&A forum.

Yes, the setup where the data source is from users' local machines and the reports are published to the Power BI service can contribute to high performance deltas and degrade performance. Here’s why:

Data Handling in Power BI

1.Import Mode:

When using Import mode, data is loaded into Power BI service memory during the refresh. This data is stored in-memory, enabling fast query performance within the Power BI service.

If the data source is from a local machine, the refresh operation needs to pull data from the local machine to the Power BI service. This can be slow and resource-intensive, especially if network latency is high or the local machine has limited resources.

2.Direct Query Mode:

In Direct Query mode, Power BI does not import the data into memory but queries the data source live for every interaction. If the data source is a local machine, every report interaction would query back to that machine, which can introduce significant latency and performance issues.

Potential Causes of High-Performance Deltas

1.Network Latency:

Pulling data from local machines to the Power BI service can be slow due to network latency.

2.Resource Contention:

Multiple reports or refresh operations competing for limited resources can lead to high memory and CPU usage, causing performance degradation.

3.Inefficient Data Refresh:

Refreshing data from local machines can be less efficient than from a centralized, optimized data source.

If you are still facing any issue, please let us know in the comments. We are glad to help you.

If the information is helpful, please Accept Answer so that it would be helpful to community members.

Thank you.