Real-time analytics services

A streaming solution follows the same shape regardless of which Azure service implements it: an event occurs, a source captures it, often a queue that keeps events in order and processes each one once, a continuously running query filters or aggregates the events, often over a time window, and the result lands in a sink such as a file, a table, or a live dashboard.

Microsoft offers several services for the processing step. Microsoft Fabric Real-Time Intelligence bundles Eventstream to ingest and route events, Eventhouse as a time series store queried with KQL, Real-Time Dashboards for live visuals, and Activator to trigger actions when conditions are met. Spark Structured Streaming is an open source library that runs on Spark based services, including both Azure Databricks and Fabric. Azure Stream Analytics is a standalone platform as a service option for defining a streaming job outside Fabric.

For the source step, Azure Event Hubs handles high volume event ingestion with ordered, at least once delivery, while Azure IoT Hub does the same job but is tuned for IoT devices specifically. Common sinks include storage such as a data lake, a database table for later querying, or Power BI for an immediate, live report.

On the exam, match IoT devices to IoT Hub, KQL to Eventhouse, and outside Fabric to Stream Analytics.