An eventhouse loading method in which data travels straight in the request body, giving close to real-time latency; it works well when lots of tables each get only small volumes. For most cases, queued (batched) ingestion remains the default because it handles more throughput.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Streaming ingestion in context, with comparison tables and the common traps.
Terms in this definition
- Eventhouse
Holds one or more KQL databases and is where Real-Time Intelligence in Microsoft Fabric keeps time-stamped streaming events for analysis, indexed by the moment they arrive.
- Get
Key Vault permission on secrets that allows a single secret to be read; App Service Key Vault references need nothing beyond it.
Related terms
- Pipeline product edition
The
editionsetting for a Lakeflow pipeline:COREcovers streaming ingestion,PROadds CDC andADVANCED, the default, adds expectations. If the pipeline uses a feature its edition lacks, the update fails. - Queued ingestion
A way of loading data into an eventhouse in which small pieces of data are grouped and then ingested together; a batch closes when it reaches a time, size or item limit (by default 5 minutes, 1 GB or 500 items, set by the ingestion batching policy). It favours throughput over latency, whereas streaming ingestion gives lower latency.