Azure Event Hub Source Node
Quick Reference
Data Asset The Azure Event Hub data asset to read from. The asset supplies the connection, Event Hub name, consumer group, and checkpoint store.
Encoding Type The format of the event bodies arriving on the hub.
Initial Position
Where to start reading when there is no prior checkpoint. Earliest replays everything retained by the hub; Latest starts from new events only.
Overview
The Azure Event Hub Source node streams events from an Azure Event Hub into your workflow. It acts as the starting point of a workflow: it joins the hub's consumer group, takes ownership of a share of the hub's partitions, decodes each event body, and passes the records downstream.
Partition ownership and read offsets are stored in a blob checkpoint store, which the data asset configures. That gives the node two properties worth knowing about: a restarted workflow resumes from where it stopped rather than re-reading the hub, and running more workers automatically redistributes partitions between them.
Connection details are not entered here — they come from the Azure Event Hub data asset you select, so the same connection can be reused across workflows without re-entering credentials.
Configuration
| Field | Description | Required | Default |
|---|---|---|---|
| Data Asset | Select or create an Azure Event Hub data asset. It provides the namespace connection or Entra ID identity, the Event Hub name, the consumer group, and the blob checkpoint container. | Yes | — |
| Encoding Type | How each event body should be interpreted. Options: CSV, JSON Object, JSON Array, JSON Lines. More info here. | Yes | JSON Object |
| Initial Position | Where to start reading on a partition that has no stored checkpoint. Earliest reads all retained events; Latest reads only events arriving after the workflow starts. Once a checkpoint exists this setting is ignored. | No | Earliest |
Notes
Give each workflow its own consumer group. Two workflows sharing a consumer group compete for the same partitions and offsets, so each ends up seeing only part of the stream. Configure a separate consumer group per independent consumer on the data asset.
Initial Position only applies the first time. After the first successful checkpoint, the workflow always resumes from its stored offset. To replay a hub from the beginning, the checkpoint data has to be cleared — changing this field alone will not do it.
Delivery is at-least-once. Events read but not yet checkpointed are redelivered after a restart, so downstream steps should tolerate the occasional duplicate.
Related Nodes
- Azure Event Hub Sink: Publish processed records to an Azure Event Hub
- Azure IoT Hub Source: Read device telemetry from Azure IoT Hub
- Kafka Source: Consume streaming data from Kafka topics