Forum Discussion

campelliann's avatar
campelliann
Icon for Post Patron rankPost Patron
3 years ago
Solved

Premium Per User VS PRO - Refresh TImes in the Service

Hi there,


I would like to know if there is a subtancial difference between Refresh times with Premium Per user and if it worth the update. I am PRO licensed now..

 

I have some datasets with highly variable refresh times (one goes from 15-40 min), and I would like to be able to reassure the user as much as possible the refresh does not fail and is timely.

It seems to me the problem is when several refreshes overlap in time. Each of the datasets on its own is relatively "fast" on the service <20 min).

  • Anonymous's avatar
    Anonymous
    3 years ago

    Hi campelliann ,

     

    The PRO license has shared capacity and the Premium Per User license has dedicated capacity.

     

    When a Pro licence schedules a refresh in Power BI Service, that refresh request enters a queue and may wait in the queue for up to an hour before being picked up and run. This is Microsoft managing the load on the Power BI servers which are used by all the Power BI community that have Pro licences. Manual refreshes are not queued but run straight away, so they appear to take less time.

     

    If the time taken to refresh is important to you, you may be able to shorten the queue wait by scheduling a refresh outside of busy business times. Scheduled refreshes at 9:00PM are often quicker than those at 9:00AM.

    Aside from that, there's not much else you can do apart from using a Premium Capacity or Premium Per User licence. The Premium licences run on dedicated capacity, so don't need to share with the rest of the community, hence queue times can be much shorter.

     

    You can refer to the following documents that may be helpful to you:

    Power BI Free vs Power BI Pro vs. Power BI Premium - Pricing & Product Comparison (folio3.com)

    Power BI Premium Per User - Power BI | Microsoft Learn

     

    Best Regards,

    Neeko Tang

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

3 Replies

  • Anonymous's avatar
    Anonymous
    Not applicable

    Hi campelliann ,

     

    The PRO license has shared capacity and the Premium Per User license has dedicated capacity.

     

    When a Pro licence schedules a refresh in Power BI Service, that refresh request enters a queue and may wait in the queue for up to an hour before being picked up and run. This is Microsoft managing the load on the Power BI servers which are used by all the Power BI community that have Pro licences. Manual refreshes are not queued but run straight away, so they appear to take less time.

     

    If the time taken to refresh is important to you, you may be able to shorten the queue wait by scheduling a refresh outside of busy business times. Scheduled refreshes at 9:00PM are often quicker than those at 9:00AM.

    Aside from that, there's not much else you can do apart from using a Premium Capacity or Premium Per User licence. The Premium licences run on dedicated capacity, so don't need to share with the rest of the community, hence queue times can be much shorter.

     

    You can refer to the following documents that may be helpful to you:

    Power BI Free vs Power BI Pro vs. Power BI Premium - Pricing & Product Comparison (folio3.com)

    Power BI Premium Per User - Power BI | Microsoft Learn

     

    Best Regards,

    Neeko Tang

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

  • bcdobbs's avatar
    bcdobbs
    Icon for Community Champion rankCommunity Champion

    From my experience if you are just talking about direct refresh of a dataset against other sources then no I don't remember seeing vast differences timing wise. Possibly a little more consistent.

     

    Where I did notice a difference was when I was doing lots of transforms in power query. Got a lot better stability and speed by using a data flow to ingest data first, doing any transforms in a second linked data flow which isn't supported in pro, before loading into dataset.