Java 17 / stream processing / resilience

Real-Time Event Stream Processor

In-memory Kafka-style event processing with partitioned brokers, cache-backed deduplication, retries, circuit breaker protection, and DLQ inspection.

100KDemo events published across partitions.
DLQFailed messages are retained for inspection.
RetryExponential backoff with jitter.

Pipeline

producer -> partitioned broker -> dedup cache -> retry handler -> circuit breaker -> success/DLQ

The Java implementation keeps every component testable without external Kafka or Redis services, which makes it useful for algorithmic system-design interviews.

Stream Simulator

4 partitionsdedup window3 retries

Run Result

Run the simulator to see delivery, duplicate, and DLQ counts.