Dataflow wont load on server but works on local - out of memory

[fth_eng]

Hello,

I have a dataset that over time my has grown with a lot of new querys (joins) to transform the data. It takes about 10 min to load the dataset on my local laptop but when I move it to the server if fails - with an out of memory error. Only way I can get it to load on the server is to switch to an A4 with 25Gb ram. On an A4 it works flawless and loads in ❤️ min. Once the datset is loaded I have good performance on the reports also on an A1 with 3Gb of ram. My laptop has 16Gb of ram but PowerBI is set to no use more than 4 Gb.

As it works locally I expect it to work also on the A1 (or at least an A2) but with a longer load time. I'm correct in this?

I'm new to PowerBi, but I'm not new to the world of ETL and BI - is there something obvious newbie settings/stuff that should check or look for?

Is there a way to analyze the performance of my datset to understand where the out of memory happens (this would also be critical if I go for a rebuild to avoid future problems)?

Thanks for your help!

Fredrik

[v-yalanwu-msft]

Hi, @fth_eng ;

Is your problem solved?? If so, Would you mind accept the helpful replies as solutions? Then we could close the thread. More people who have the same requirement will find the solution quickly and benefit here. Thank you.

Best Regards,

Community Support Team_ Yalan Wu

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[v-yalanwu-msft]

Hi, @fth_eng ;

It's recommended that you have at least two times more space available than your dataset for refresh.

https://docs.microsoft.com/en-us/power-bi/admin/service-admin-premium-workloads

https://community.powerbi.com/t5/Service/Dataflow-to-Dataset-refresh-memory-usage/td-p/1023369

https://www.element61.be/en/resource/microsoft-power-bi-performance-tracing-power-bi-premium

Best Regards,
Community Support Team_ Yalan Wu
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.