Discrete events or a continuous stream?
Both services move events, but they solve different problems. Choosing the right one keeps costs down and latency predictable.
Push one event, or stream millions
Event Grid
Discrete events, pushed to subscribers
BlobCreated
subject: /invoices/2026/inv-881.pdf
↻ retry with exponential back-off
Event Hubs
A continuous stream, split into partitions
Events from the same device always land in the same partition, so their order is preserved.
Illustrative example
Azure Event Grid
Event routing service for building event-driven architectures. Event Grid provides:
- Reactive programming: React to Azure resource state changes
- Custom topics: Publish application-specific events
- Event filtering: Advanced filtering and routing logic
- High throughput: Up to 5,000 events per second per custom topic, or per throughput unit on a namespace
Azure Event Hubs
Big data streaming platform for high-throughput event ingestion. Event Hubs enables:
- Telemetry ingestion: IoT devices, application logs, clickstream data
- Kafka compatibility: Works with existing Kafka clients
- Partitioning: Parallel processing for scalability
- Capture: Automatic archival to Azure Storage or Data Lake