I want to send streaming data to Power BI for reporting purposes. What should I take into account when choosing the right architecture?
 |
| Streaming Data to Power BI |
Solution
If you for example have a helpdesk for your customers where they can call or chat for support then you probably also want some real time reports to see the current state of the calls and chats. Most regular Data Warehouses are often only refreshed once a night and then it's already too late to react to incidents.
For real time reports in Power BI have two main options. The first option is to send the events directly to a Power BI Streaming dataset (Push or Streaming) and then build a report and pin reports visuals to a dashboard. This is an appropriate solution for a lot of real time reports, but there are some limitations. For example there is a maximum number of events per second. Once you exceed that limit you start loosing data. Propably just when you need accurate reports the most: when it is very busy in your helpdesk. An other limitation for streaming datasets in Power BI is the history. It keeps only one hour of data.
The second option is to push the data into Azure Event Hubs and then use Azure Stream Analytics to push it to Power BI. This solves the max number of events per second because Stream Analytics can aggregate or filter the data before sending it to Power BI and Stream Analytics can also send it to for example a data lake to solve your history problem.
 |
| Streaming Data to Power BI |
In this streaming data series we will explain this second option focussing on the hot path and the capture in the data lake which is part of the cold path. Just like for a 'regular' data warehouse architecture there are a lot of different solutions, but this one is probably the most common and simple solution that will fit the majority of cases. One particular new streaming data feature in Azure that is worth mentioning, is writing to a Delta Lake table. At the moment of writing this is still in public preview and only available in a limited number of Azure regions, but this will fit the Lake House architecture very well.
Posts in this series:
- Setting up Azure Event Hubs
- Sending test messages to the Event Hub
- Setting up hot path in Azure Stream Analytics
- Setting up cold pathin Azure Stream Analytics *soon
- Viewing streaming data in Power BI
