🚧 Under construction — some information may be outdated.
Projects

Developer Nov 2025

A reverse proxy and load balancer built to distribute HTTP traffic across a cluster of backend servers, with a focus on concurrency control, fault tolerance, and observability.

I implemented two routing strategies — Weighted Round Robin and Least Connections — backed by hybrid health checks: active TCP pings on a configurable interval, and a passive circuit breaker that instantly removes nodes spiking 5xx errors. Automatic failover retries make individual backend failures transparent to the client.

On the resilience and security side, I built TLS termination to offload SSL processing at the entry point, and a per-IP Token Bucket rate limiter to guard against DDoS and noisy-neighbor exhaustion. A live dashboard visualizes connection pools, error rates, and server status in real time. All shared state is managed with sync/atomic and sync.RWMutex for safe concurrent access.

Fulcrum live dashboard