Modern data platforms must ingest heterogeneous real-time data from multiple producers—each sending different schemas, versions, or malformed messages.

In this technical session, we’ll explore how Microsoft Fabric handles schema variability using schema inference, schema sets, and dynamic routing inside Real-Time Intelligence.

You’ll learn practical patterns for identifying incoming schemas, processing each one independently, detecting “bad messages,” and building a robust ingestion pipeline.

We will also demonstrate the new HTTP Endpoint feature for low-latency ingestion outside other traditional flows.