Power BI limitations for large datasets

[alvin3828]

If someone is using Power BI connected to Azure as a front-end for data visualization. 

What are some limitations? 

In the past, even with smaller data sets in SQL, some charts took 25-30 seconds to render? 

What I am trying to answer is at what point does the visualization take very long based on: 

1. Size of the dataset 

2. Number of different datasets or types queried (multiple sql + no sql databases)

3. Complexity of analysis or statistics performed (writing R code to regression & other complex statistical calculations, clustering etc) 

I am trying to understand the limiations of Power BI as a front-end before deciding to develop a custom front end or use another off-the shelf BI took. 

I would likely be using Power BI connected to an Azure stack (SQL server, documentDB, etc). 

[Eric_Zhang · Accepted solution]

@alvin3828

The listed 3 points would more or less have affect of performance per my experience, however I've never experienced a perfomance as slow as 25-30 seconds. It usually doesn't take that long when loading a report.

One known fact that impacts  the performance a lot is using a Direct Query to Azure SQL DB or on-premise database, the query would run for long if the data in DB is large.

By the way, you could try to debug the performance by using the brower dev tool(F12 in Chrome,IE,FF) to find what network traffic consume long time.