Graceful Shutdown Patterns in Golang Services

Production services must handle SIGTERM gracefully. Kubernetes sends SIGTERM before killing pods — you have ~30 seconds to finish in-flight work.

The Pattern

func main() {
    srv := &http.Server{Addr: ":8080", Handler: router}
    
    go func() {
        if err := srv.ListenAndServe(); err != http.ErrServerClosed {
            log.Fatal(err)
        }
    }()

    quit := make(chan os.Signal, 1)
    signal.Notify(quit, syscall.SIGINT, syscall.SIGTERM)
    <-quit

    ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
    defer cancel()

    if err := srv.Shutdown(ctx); err != nil {
        log.Fatal("forced shutdown:", err)
    }
}

What to Close

During shutdown, ensure you close:

1. HTTP server (drain in-flight requests)
2. Database connection pools
3. Redis connections
4. Background worker goroutines
5. Message queue consumers

Health Checks

Configure readiness probes to fail immediately on SIGTERM. This removes the pod from the load balancer before shutdown begins, preventing new requests from arriving.

Key Takeaways

• Always use http.Server.Shutdown() with a timeout
• Close database pools explicitly
• Fail readiness probes on shutdown signal
• Log shutdown duration for observability