Sentrinet
Distributed Network Intelligence Platform
Problem
Network security teams need real-time visibility across infrastructure at scale. Sequential scanners create bottlenecks that make large-scale monitoring impractical.
Architecture
Go-based distributed scanner using goroutine worker pools with a task scheduler that orchestrates long-running monitoring workloads. PostgreSQL for persistent state, Redis for caching, and Prometheus for real-time observability.
Engineering Highlights
- ▸Executes 1000+ concurrent network operations using goroutines, channels, and bounded worker pools
- ▸Fault-tolerant scheduler with retries, timeout handling, and asynchronous execution
- ▸REST APIs with Prometheus monitoring for production-grade observability
- ▸Scalable backend architecture emphasizing concurrency and maintainability
Tradeoffs
Chose Go over Rust for faster iteration and goroutine ergonomics. PostgreSQL over time-series DBs for query flexibility. Bounded worker pools prevent resource exhaustion while maintaining throughput.