Forum Discussion
Choosing the right Power Bi Premium server
Hi all,
I need some advices on the right Power Bi Premium server package to choose. I have a huge dataset and I'm afraid the power Bi premium pricing calculator do not reflect my reality.
I have only 300 users, 10 pro users. According to the calculator, only one P1 nodes is necessary.
However, one table of my data model has 1.6 billion data. Another one has 200 millions. See data model attached.
Should we plan for more capacity and fees? How much?
Thanks,
Joel
https://drive.google.com/file/d/1-HPVkskITfGvOZ4476GMcQVXfVncpAuP/view?usp=sharing
- Hi there,
It would appear that you should be ok on a P1 with incremental refreshing.
There is a new premium capacity metrics app which can be used once implemented to see how everything is performing.
And as I'm sure you have read before if there is anything to make the model smaller look into that. Like unused columns and removing the built in date columns.
5 Replies
- GilbertQ
Super User
Hi there
How large is your PBIX file with all the data loaded?
Are you looking to use Row Level Security?
The above will give an initial indication in terms of how much capacity you require.
It should be noted that you will need additional capacity for the refreshing of the data, but you can use incremental refreshing to ensure that you do not need double the capacity to refresh the data.Hi GilbertQ,
thanks for your insights.
The PBIX file is 8Gb.
There is 4 level of security applied using RLS : Salespersons (250), manager (30); director (10) and admin (10)
We plan to use the incremental refreshing functionnality.
Base on this information, have you got any other insights?
thanks,
Joel
- GilbertQHi there,
It would appear that you should be ok on a P1 with incremental refreshing.
There is a new premium capacity metrics app which can be used once implemented to see how everything is performing.
And as I'm sure you have read before if there is anything to make the model smaller look into that. Like unused columns and removing the built in date columns.