Forum Discussion

Mattw112's avatar
Mattw112
Icon for Helper I rankHelper I
2 years ago

Capacity and Planning Online

We have Premium Capacity P2 and pretty robust Virtual machines hosting the gateway server.  P2 averages around 50% capacity.   I have a group that uses this capacity that wants to share reports wit...
Need Help
TomMartens's avatar
TomMartens
Icon for Super User rankSuper User
2 years ago

Hey Mattw112 ,

 

2000 additional users will have a great effect on the capacity resources.

More users means more CPU utilization, of coures more RAM is required as well. If RLS is involved with a lot us different roles, than RAM can become pressured, especially if it's dynamic RLS.

Unfortunately, there is no golden rule, if these 22k users are spread across many different time zones, then it can also be likely that you will not experience a high utilization of the capacity resources.

 

I consider a the resources management of Power BI capacity very efficient, aged chunks of data are evicted from RAM fast freeing up memory for new requests (e.g., changing a slicer selection), but data will be loaded to memory very fast.

 

Nevertheless, efficient DAX becomes very important if many users are hitting the same report at the same time, if RLS is involved, you might detect a measure of death. 

 

Get familiar with Premium capacity metrics app, watch it closely 🙂

 

2k new users is a large success regarding Power BI adoption! Congratulations!

 

Hopefully, this helps somehow.

 

Regards,

Tom