EventRouter, Workflows, and Data Streams comparison
Written by
Updated at August 12, 2025
| Criterion \ service | EventRouter | Workflows | Data Streams |
|---|---|---|---|
| Architecture approaches | Choreography, data bus. Independent integrations with other services and message processing, quick change of message processing settings |
Orchestration. Sequential execution of interconnected workflow steps |
Message broker |
| What it can help with | Delivery, routing, filtering, and transforming messages | Filtering and processing messages, integration across services, and sequentially executing steps | Storing messages, scaling in case of loads |
| Restrictions | Message delivery timeout, number of message targets, message size | Workflow timeout, message size | Message retention period |
| Streaming message processing | Yes | No | No |
| Batch message processing | Yes | No | No |
| Sending messages from multiple sources to multiple targets | Yes | Can be implemented in custom code | Yes |
| Interface | Rules that define where messages are delivered to, how they are processed, and where they are redirected to. | Visual constructor for designing workflows or a YAML specification. | Amazon Kinesis Data Streams API, AWS CLI, Kafka API (with limitations) |
| Error handling | You can redirect unprocessed messages to a dead letter queue; processing retry mechanisms are supported. | Compatible with processing retry mechanisms and the error handling logic. | No |
| Monitoring and logging | Compatible with monitoring and logging for message transmission over a bus. | Compatible with monitoring and logging for step execution with a visual timescale. | Compatible with monitoring and logging for message transmission over a data stream. |