Power BI Premium Gen 2 First Look

Currently Premium Gen 2 is in Preview (Jan 20201) but it looks really exciting for those with Premium capacity. Lets have a look at Gen 2 and see how it differs to Gen 1

Gen 1

With Premium Gen 1 we are bound by the number of vCores and by the memory we have. Lets take the P1 SKU as an example

P1, P2 P3 P4 = Are all Premium SKUs. Each one has more vCores for better performance

SKU = Stock Keeping Unit

vCores = Virtual Cores is a puchasing model which gives you more control over compute and memory requirements.

When it comes to Premium, reports would slow down if there were two many queries running. And if people were using many models at the same time you would have to wait for your time slot.

As you can see, there is only 25 gig memory across the data sets Once a report isn’t being run and used any more, that memory gets dropped and is added back into the pool.

Report users and report developers are also fighting with report refreshes.

Premium Gen 2

Gen 2 is not in generally availability yet but if you have Premium you can switch to Gen 2 in preview.

In Power BI Admin (as the Power BI administrator or Global Admin)

Go to the Capacity Settings and switch Gen 2 from disabled to enabled.

You can also go back to Gen 1 if you need to but if you do make sure you flag up any issues you are having.

lets have a look at the model compared to Gen 1

Autoscaling

The end users encounter the throttling and performance issues with Gen1 because they physically only have 4 backend vCores. Now with Gen2, Autoscaling allows you to deal with spikes. This is not available yet but will be coming. This is helped by the fact that there are other vCores that can be called on.

If you do come up to the 4 core Limitation it will, or may lend you a vCore so you don’t see impact for your end users

Previously our admins had to deal with this kind of problem but this will really help automate these kind of issues

Memory

Data Sets can go over the 25 gig memory capacity. Previously Premium was 25 gig for all the data sets. Now Data sets are gated individually.

This is a fantastic update. We don’t have to worry about the collective size of all our data sets.

Refreshes

Previously there was a maximum of 6 refreshes at any one time. Otherwise you can get throttled.

With Gen 2, refreshes get spread out over a 24 hour period and don’t impact other queries from users. refreshes just run

The looks great. People are seeing refreshes of an hour and a half coming down to 10 minutes.

Capacity Usage Metrics

This is coming soon and will have a breakdown of items.

Its a little annoying when you have set up gen 2 and want to view the metrics to see how everything is working but currently can’t.

With Gen 2 we will also be able to work against a chargeback model. This means that we can spread the costs of Premium between distinct areas of an organisation dependant upon their usage.

Workloads

Again the workload settings aren’t fully functional at the moment but more will be coming.

For example for data set workloads we can specify minimun refresh intervals and execution intervals. we can detect changes in our metrics.

We don’t have settings for dataflows and AI yet.

Why go with Premium Gen 2 Checklist

• Performance benefits
• End users see faster reports
• Refreshes, we now don’t have refresh bottlenecks and we remove refresh failures due to throttling
• Premium per user
• Improved metrics will be introduced soon
• Autoscaling
• Proactive admin notifications

Why it may be worth waiting until Preview becomes GA

It looks like people are having some issues with dataflows and there is already be a known issue about this

It looks like this might be fixed quickly, and once dataflows are OK, it seems like. A workaround is to move your dataflows out into another Workspace and then back in but hopefully this will get much better.

Questions

Is Premium Gen 2 going to be the same price as Gen 1?

Is there any way to find out how many dataflows you have if dataflows are an issue?

Will we still give great functionality to the Power BI Pro users?