Data capacity

[douglasday]

I want to understand the Power BI Embedded capacity model in more detail.  If anyone can anaswer these questsion it would be greatly appreciated.

Question

I have a PBI workbook, the file size is 1.5GB on disk and when opened on PBI dekstop it consumers around 500MB of memory,  When I upload this to our Power BI Embedded environment (running A1).  The workbook consumes 1.5 GB of resources.

1) Why does it consume so much memory when on PBI dekstop it only uses 500MB, surely it cant be the file size!

2) Can I control when that memory is actually utilised.  For example if no users are actively looking at the report shouldn't the report be unloaded and not consuming any capacity?

Regards

[v-micsh-msft]

So the Excel data is imported into Power BI Desktop?

How many DataSets are involved under your Power BI Embedded report?

1. The Memory usage is an average data, you may take a look at the following article about the metrics:

Metric Description

2. Yes, you could pause the capacity through the Power BI Embedded API, or through the Azure portal, when there is no report refresh requests:

Pause and start your Power BI Embedded capacity in the Azure portal

Regards,

Michael

[douglasday]

Regarding managing capacity. 

I have multiple workbooks loaded to azure.  each workbook represent a single customer.  I was expecting that capacity (memory) would only be consumed when a user was actually accessing the workbooks, not loaded into memeory all the time.  I'd at least expect the ability to load and unload a workbook from memory to control the utilisationj of my capcity!

I dont mind creating seperate capcity for each client, with a workbook in each and starting and stoping the service.  The problem is the services take 60-90 seconds to start up.  Which is way to long for a web application.  Do you know of a way to improve this startup time?