Embedded Monitoring Best Practice: Use go-commons for Built-in Service Health Reports

Small teams and individual developers often face three monitoring problems: the heavyweight node_exporter requires a separate deployment; viewing only CPU and memory forces the setup of a full Prometheus + Grafana stack; and business code and system metrics are scattered, preventing a unified view.

Quick installation

go get github.com/Rodert/go-commons

Minimal monitoring service

The following Go program (<50 lines) exposes a business metric (QPS) together with system metrics (memory usage and CPU usage) on a single /metrics endpoint.

package main

import (
    "fmt"
    "net/http"
    "sync/atomic"
    "time"

    "github.com/Rodert/go-commons/metrics"
)

var qpsCounter int64

func main() {
    // Simulated business handler
    http.HandleFunc("/hello", func(w http.ResponseWriter, r *http.Request) {
        atomic.AddInt64(&qpsCounter, 1)
        fmt.Fprintln(w, "Hello, go-commons!")
    })

    // Metrics endpoint
    http.HandleFunc("/metrics", func(w http.ResponseWriter, r *http.Request) {
        mem := metrics.GetMemoryUsage() // memory usage
        cpu := metrics.GetCPUUsage()    // CPU usage
        qps := atomic.SwapInt64(&qpsCounter, 0)

        fmt.Fprintf(w, "qps %d
", qps)
        fmt.Fprintf(w, "memory_usage %.2f
", mem)
        fmt.Fprintf(w, "cpu_usage %.2f
", cpu)
    })

    // Reset QPS every second
    go func() {
        for range time.Tick(time.Second) {
            atomic.StoreInt64(&qpsCounter, 0)
        }
    }()

    fmt.Println("server started at :8080")
    http.ListenAndServe(":8080", nil)
}

Run the service

go run main.go

Verify the monitoring output

Call the business endpoint: curl http://localhost:8080/hello Fetch the metrics: curl http://localhost:8080/metrics Sample output:

qps 3
memory_usage 42.78
cpu_usage 5.13

This behaves like a miniature node_exporter , but the metrics are integrated directly with the application, requiring no separate deployment.

Next step: integrate with Prometheus + Grafana

Add a job to the Prometheus configuration:

scrape_configs:
  - job_name: "go-app"
    static_configs:
      - targets: ["localhost:8080"]

Grafana can then display a dashboard that combines QPS and system metrics.

Why go‑commons is recommended

Zero‑cost integration: a single go get and a few lines of code.

Lightweight: no need to deploy a separate node_exporter.

Unified view: business and system metrics are served from the same /metrics endpoint.

Open‑source and actively developed; future versions may add network, disk, and goroutine metrics.

Contribute

The project is early‑stage and welcomes feedback, code contributions, and documentation improvements. Repository: https://github.com/Rodert/go-commons