Forum Discussion
Composite mode in PBI
- Is it possible to combine visuals from both datasets (streamed + live query) in a single Power BI dashboard?
- If yes, are there any limitations I should be aware of (e.g., refresh behavior, interactivity)?
Use case:
Consumption of real time streaming data from an existing Kafka topic in Power BI. The data should refresh in real time on the report.
Architecture:
- A small Kafka consumer app( Python/ Java )reads from Kafka.
- That app pushes data into Power BI (using Power BI Push Dataset API)
- Power BI visualizes it live, auto refreshing as new data comes in.
- Push Dataset: Create a dataset in Power BI that accepts incoming data via API.
- Kafka consumer: Read the messages continuously from Kafka.
- Small App: Takes Kafka data , flattens the JSON and sends it to Power BI using REST API.
- Auto refresh: Power BI dashboards automatically refresh when new rows are pushed.
Hi Anonymous ,
Thank you for the follow-up!
Since you are on Power BI Premium P1 without Fabric Capacity, features like Eventstreams and KQL databases unfortunately are not available .They require Microsoft Fabric .
Another option is Streaming datasets through Azure Stream Analytics, which might also feed live data into dashboards, though with limited query flexibility.
Refer the below links for more information:
https://learn.microsoft.com/en-us/azure/stream-analytics/stream-analytics-real-time-fraud-detection
https://learn.microsoft.com/en-us/power-bi/connect-data/service-real-time-streaming
Hope this resolve your query.If so,give us kudos and consider accepting it as solution.
12 Replies
- Deku
Super User
Wouldn't advise push. Why not just push Kafka to a KQL database using a eventstrem and use direct query to pull into power bi? You can setup a refresh interval on the power bi page to pull fresh data every second, 10 sec ...
In regard to composite, it will all depend on the cardinality of the relationship, as relationships will be respected by a large where clause to the remote source
Hi Anonymous ,
Thank you Deku for the helpful response!Yes, if you decide to go with the KQL (Kusto Query Language) + Eventstream route in Fabric, then Fabric Capacity is required, since these features are part of the Real-Time Analytics workload. You may set up an Eventstream to pull data from your Kafka topic and write it directly into a KQL database, which Power BI can then query using DirectQuery. This allows for near real-time dashboards, with auto-refresh intervals as low as 1 second.
While this setup offers more flexibility and scalability than Push datasets, it’s currently Azure-native, so Fabric is the main path if you want tight integration with Power BI.
Please refer the links below for more information:
https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-composite-models
https://learn.microsoft.com/en-us/rest/api/power-bi/push-datasets/datasets-post-rows
https://learn.microsoft.com/en-us/power-bi/connect-data/service-real-time-streaming
I hope this resolve your query.If so,give us kudos and consider accepting it as solution.
Regards,
Pallavi.
