Designing architecture for high concurrency: Stateless horizontally scalable WebSocket gateways Redis Pub/Sub for distributed event propagation Connection sharding strategy Message batching to reduce fan-out overhead Backpressure handling for slow consumers Binary payload optimization p95/p99 latency monitoring Rate-limited presence updates Async persistence queue for chat history Optimizing for: Low memory footprint per socket Event-loop safety Minimal serialization overhead Distributed scalability Generating production-grade implementation with: Redis adapter Connection pooling Streaming architecture Horizontal autoscaling support Benchmark-ready instrumentation