Architecture and Price estimate?

[Skydancer]

Hello.

Let's say I want to provide SAAS reports to 1,000 customers, gathering data from every customer individually. Meaning 1 customer = 1 company with, different data sources with different internal structure, with different type (can be SQL, can be API from cloud).  I want to have couple of standardizes reports which cusomer presellect. Every customer have approximately 1,000,000 - 5,000,000 fact table rows and 2,000 dimension rows. 

Data gathered is of the same type (sales, employees, etc, dim tables). But, because of different internal structure need to be structured and prepared (ELT).

I would expect every customer to spend about 2 to 10 minutes browsing the report every day. 

If there is anyone with experience of similar scenario implementation or usage, can you please recommend first of all, the general level infrastructure? Like, for example, data lake -> data warehouse -> Analysis Server -> Power BI Embeded. And secondly, what is the rough (even if VERY rough) Azure infrastructure cost per user per month or another costing understarding?

[Greg_Deckler · Accepted solution]

@Skydancer - I think you have the basic components correct. The price is going to vary wildly depending upon how much transformation you need to do, where you do it and what tool you use. For example, Databricks is wildly more costly than Azure Data Factory. Same goes for your data warehouse, are you using Azure Data Warehouse or Azure SQL DB or Azure SQL Managed Instance? Are you going to use Premium or Standard for this stuff? Huge price differences. LRS, ZRS or GRS? Are you including dev and test environments? 

Here is an example link: https://azure.com/e/e1020c8e0b034622a82b4ac17eff5ac1

[Skydancer]

Thank you Greg.

It gives rough understanding of the cost.

Still I am wondireing if the situation described is so rare, that there are no other comments from actual users. 

P.S. I just bought your book 🙂