Real-Time Intelligence is the workload inside Microsoft Fabric for data that needs to be queried and acted on as it arrives, rather than after a scheduled refresh. It ingests streaming data, stores it in a query engine built for time-series data, and can trigger an alert or an action the moment a condition is met.

The components

Eventstream sources include Azure Event Hubs, Azure IoT Hub, Kafka clusters, and change data capture feeds from a database. A single eventstream can route the same data to an Eventhouse for querying and to a Fabric Lakehouse for batch use at the same time.

A worked example

A common pattern: monitoring a stream of sensor readings and alerting when a value crosses a threshold.

1. Create an Eventstream and connect it to the source, for example an IoT hub emitting temperature readings.

2. Set the Eventstream's destination to a new Eventhouse, which creates a KQL database to hold the incoming events.

3. Open a KQL queryset against that database to confirm data is arriving and shape a query:

   SensorReadings
   | where Timestamp > ago(1h)
   | where Temperature > 80
   | summarize MaxTemperature = max(Temperature) by SensorId, bin(Timestamp, 5m)

4. Build an Activator rule on the same Eventhouse table that fires when Temperature exceeds 80, sending an alert to a Teams channel or an email address.

5. Optionally, add a Real-Time dashboard on the Eventhouse for a live view, or connect Power BI to the same data for reporting alongside the rest of your semantic model. See Microsoft Fabric and Power BI for how a Power BI report reads Fabric data.

Nothing in this flow runs on a schedule. The Eventhouse ingests continuously, and the Activator rule evaluates as new rows arrive.

When Real-Time Intelligence fits

Real-Time Intelligence fits scenarios where the delay between an event happening and a report showing it matters: fraud detection, equipment monitoring, operational alerting, or any dashboard that needs to reflect what is happening now rather than what happened as of last night's refresh. A daily sales summary or a monthly finance report does not need it. A standard Data Factory pipeline into a Lakehouse or Warehouse, refreshed on a schedule, is the simpler and cheaper choice for anything that does not need to react within seconds. For the rest of what Fabric includes beyond real-time data, see our Microsoft Fabric overview.

FAQs

What is Real-Time Intelligence in Microsoft Fabric?

The Fabric workload for ingesting, storing, querying and acting on streaming data as it arrives, built from Eventstream, Eventhouse, KQL querysets, Activator and Real-Time dashboards.

What is the difference between Eventstream and Eventhouse?

Eventstream moves and transforms data in flight, routing it to one or more destinations. Eventhouse is where that data lands and gets stored for querying. A single eventstream typically feeds one or more eventhouses.

Do I need to write code to use Real-Time Intelligence?

Eventstream and Activator are both no-code, built around a visual canvas and rule builder. Querying an Eventhouse benefits from knowing KQL, though the same data can also be queried in T-SQL through a KQL queryset.

Sources

• What is Real-Time Intelligence in Microsoft Fabric? - Microsoft Learn

• Eventhouse overview - Microsoft Learn