Forum Discussion
Automated PDF Generation
You actually purchase the A SKUs through the Azure portal. see https://learn.microsoft.com/en-us/power-bi/developer/embedded/azure-pbie-create-capacity?tabs=portal
And there are details here on how to pause/start an embedded capacity https://learn.microsoft.com/en-us/power-bi/developer/embedded/azure-pbie-pause-start with links to ways to automate this.
Once you have created and started the embedded capacity you then need to go back to Power BI and into the setting for your workspace in the Premium setting set it to use this A SKU. Note that once you do this the embedded capacity needs to be running for the reports to work.
The standard P SKUs are really aimed at large enterprises with 500 or more report viewers. For smaller organizations with a monthly report burst like this the A SKU should be more cost effective. The note here https://learn.microsoft.com/en-au/power-bi/enterprise/service-admin-premium-testing#purchase-a-skus-for-testing-and-other-scenarios says that the A SKU gives you all the features of a standard premium capacity, but without the unlimited sharing (so all your users will require Pro or PPU licenses, but it sounds like you have these already)
We need to do hundreds, so while we could add a delay or even add several PPU accounts to increase capacity, it would not be the simple and scalable solution we're looking for.
Thanks,
Chris
marquis_chris wrote:
We need to do hundreds, so while we could add a delay or even add several PPU accounts to increase capacity, it would not be the simple and scalable solution we're looking for.
So the other option inbetween a PPU and P1 SKU would be to look at using an A SKU. https://learn.microsoft.com/en-us/power-bi/developer/embedded/embedded-capacity#sku-memory-and-computing-power but to make this cost effective you would also need to look at either scaling up when you run your exports and then scaling down or pause the capacity when it is not being used. If you were to run an equivalent A SKU 24x7 it would cost more than the same size P SKU.
I would suggest you start with a small scale trial, run a small sample of your reports and see how long they take and use the Premium Capacity Metrics App (https://learn.microsoft.com/en-us/power-bi/enterprise/service-premium-install-app ) to see what the CPU usage is like. You should then be able to project roughly what the CPU usage will be like for your full run. This article has more information on capacity planning for paginated reports https://learn.microsoft.com/en-us/power-bi/paginated-reports/paginated-capacity-planning If you run your test over an hour or two it should be pretty cheap since A SKUs are billed hourly and running an actual sample work load is the best way to get a good estimate of your full workload.