#Goroutine

I build backend systems for a living. Most of what I ship runs on NestJS — TypeScript, async/await,...

Go channels are synchronization primitives, not queues. They deliver backpressure only when the producer is bounded — and that's where the real OOM hides.

The race detector finds bugs nothing else will. It also costs 2-20x CPU per the Go team's own docs. Three sampling strategies to keep it on in prod.

A parked goroutine burns zero CPU, so flame graphs miss leaks entirely. Read goroutine.pprof, set a NumGoroutine baseline, and gate CI with goleak.

From Dev.to - grpc: Postmortem: How a Go 1.24.0 gRPC 1.60.0 Deadlock Caused a 4-Hour Outage for 2026 FinTech App