Stream data to millions

Reference Architecture

Zilla enables scalable, real-time distribution of Kafka data streams to browsers and mobile clients by translating Kafka streams into Server-Sent Events (SSE). Since Kafka isn't built to handle a high volume of client connections, Zilla efficiently manages this by off-loading client connections through an internal cache. This approach maintains a minimal number of direct connections with Kafka brokers, regardless of how many clients are simultaneously connected.
‍
Besides SSE fan-out, Zilla also supports turning Kafka into a fully-fledged REST web server through which clients can GET, POST, PUT, and DELETE data streams.
‍
Zilla is stateless, therefore easy to scale. It can be deployed behind a load balancer and run as as a standalone process or as part of a Kubernetes cluster. For observability, Zilla exposes metrics via OpenTelemetry, while message validation is supported via AsyncAPI and Schema Registry.

Feature Highlights

Server Sent Events (SSE)

SSE is an HTTP-based server push protocol. It's part of HTLM5, so is natively supported all modern browsers and mobile frameworks.

Fan-Out with Message Filtering

Zilla supports fan-out with filtering by message key and header, enabling different clients to subscribe to individual SSE streams.

Reliable Delivery

Zilla sends an event id with every message. A client can send a last-event-id header to recover from an interrupted stream without message loss.

Continous JWT Authorization

Zilla continuously authorizes SSE streams by sending a "challenge" event, triggering the client to send up-to-date authorization credentials before expiration.

App-centric REST APIs

Establish an application-centric, idempotent REST APIs that allows clients to GET, POST, PUT, DELETE Kafka message streams.

OpenAPI and AsyncAPI

Use OpenAPI and AsyncAPI specs for configuration and message validation in Zilla. Augment an existing OpenAPI by mapping it to Kafka and adding SSE endpoints for push updates.