Forum Discussion
How to get additional resources (Memory ) for data flows to run faster ?
Hi Community,
We have been struggling to manage datasets and data flows refreshes in one Premium capacity which have P3 (32 cores) . Many times, datasets and data flows(backend operations) are not finding resources because of multiple user interactions on reports available in Service.
We are planing to scale up our capacity and dedicate seperate memory and make the resources available for data flows and datsets to run faster ( refreshes) . Is there a way to do this ? Please assist !!!!
G Venkatesh
Hi Anonymous
For the Linked entities to refresh automatically you will need to keep them in the same workspace, please refer to the article below.
https://docs.microsoft.com/en-us/power-bi/service-dataflows-linked-entities
You first entity should be simple with minor transformations, just to pull the data from your data source ( Ideally with incremental refresh ), the second ( and third if applicable ), should be where you do most of your transformation and later use to build reports.
Best Regards,
Mariusz
If this post helps, then please consider Accepting it as the solution.
Please feel free to connect with me.
LinkedIn
8 Replies
- MariuszCommunity Champion
Hi Anonymous
You can optimize your dataflows, allocate max resources from your P3 (if you haven't it's up to 20% )
https://docs.microsoft.com/en-us/power-bi/service-dataflows-overview
https://docs.microsoft.com/en-us/power-bi/service-admin-premium-workloads
Get a separate instance of azure data lake gen2 and do your transformation there ( a lot cheaper than premium nodes not restricted by 20% ).
Best Regards,
Mariusz
If this post helps, then please consider Accepting it as the solution.
Please feel free to connect with me.
LinkedIn- AnonymousNot applicable
Hi Mariusz,
Thanks for the reply. Unfortunately, we do not have our Own ADLF in my organization . As of now, below is the work load pattern for data flows we have set. Do you recommend any changes ?
Kindly let me know your inputs.
G Venkatesh