Go指标评审

prometheus-go-code-review

by anderskev

Reviews Prometheus instrumentation in Go code for proper metric types, labels, and patterns. Use when reviewing code with prometheus/client_golang metrics.

3.7k编码与调试未扫描2026年3月23日

安装

claude skill add --url github.com/openclaw/skills/tree/main/skills/anderskev/prometheus-go-code-review

文档

Prometheus Go Code Review

Review Checklist

  • Metric types match measurement semantics (Counter/Gauge/Histogram)
  • Labels have low cardinality (no user IDs, timestamps, paths)
  • Metric names follow conventions (snake_case, unit suffix)
  • Histograms use appropriate bucket boundaries
  • Metrics registered once, not per-request
  • Collectors don't panic on race conditions
  • /metrics endpoint exposed and accessible

Metric Type Selection

MeasurementTypeExample
Requests processedCounterrequests_total
Items in queueGaugequeue_length
Request durationHistogramrequest_duration_seconds
Concurrent connectionsGaugeactive_connections
Errors since startCountererrors_total
Memory usageGaugememory_bytes

Critical Anti-Patterns

1. High Cardinality Labels

go
// BAD - unique per user/request
counter := promauto.NewCounterVec(
    prometheus.CounterOpts{Name: "requests_total"},
    []string{"user_id", "path"},  // millions of series!
)
counter.WithLabelValues(userID, request.URL.Path).Inc()

// GOOD - bounded label values
counter := promauto.NewCounterVec(
    prometheus.CounterOpts{Name: "requests_total"},
    []string{"method", "status_code"},  // <100 series
)
counter.WithLabelValues(r.Method, statusCode).Inc()

2. Wrong Metric Type

go
// BAD - using gauge for monotonic value
requestCount := promauto.NewGauge(prometheus.GaugeOpts{
    Name: "http_requests",
})
requestCount.Inc()  // should be Counter!

// GOOD
requestCount := promauto.NewCounter(prometheus.CounterOpts{
    Name: "http_requests_total",
})
requestCount.Inc()

3. Registering Per-Request

go
// BAD - new metric per request
func handler(w http.ResponseWriter, r *http.Request) {
    counter := prometheus.NewCounter(...)  // creates new each time!
    prometheus.MustRegister(counter)       // panics on duplicate!
}

// GOOD - register once
var requestCounter = promauto.NewCounter(prometheus.CounterOpts{
    Name: "http_requests_total",
})

func handler(w http.ResponseWriter, r *http.Request) {
    requestCounter.Inc()
}

4. Missing Unit Suffix

go
// BAD
duration := promauto.NewHistogram(prometheus.HistogramOpts{
    Name: "request_duration",  // no unit!
})

// GOOD
duration := promauto.NewHistogram(prometheus.HistogramOpts{
    Name: "request_duration_seconds",  // unit in name
})

Good Patterns

Metric Definition

go
var (
    httpRequests = promauto.NewCounterVec(
        prometheus.CounterOpts{
            Namespace: "myapp",
            Subsystem: "http",
            Name:      "requests_total",
            Help:      "Total HTTP requests processed",
        },
        []string{"method", "status"},
    )

    httpDuration = promauto.NewHistogramVec(
        prometheus.HistogramOpts{
            Namespace: "myapp",
            Subsystem: "http",
            Name:      "request_duration_seconds",
            Help:      "HTTP request latencies",
            Buckets:   []float64{.005, .01, .025, .05, .1, .25, .5, 1, 2.5, 5, 10},
        },
        []string{"method"},
    )
)

Middleware Pattern

go
func metricsMiddleware(next http.Handler) http.Handler {
    return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
        timer := prometheus.NewTimer(httpDuration.WithLabelValues(r.Method))
        defer timer.ObserveDuration()

        wrapped := &responseWriter{ResponseWriter: w, status: 200}
        next.ServeHTTP(wrapped, r)

        httpRequests.WithLabelValues(r.Method, strconv.Itoa(wrapped.status)).Inc()
    })
}

Exposing Metrics

go
import "github.com/prometheus/client_golang/prometheus/promhttp"

func main() {
    http.Handle("/metrics", promhttp.Handler())
    http.ListenAndServe(":9090", nil)
}

Review Questions

  1. Are metric types correct (Counter vs Gauge vs Histogram)?
  2. Are label values bounded (no UUIDs, timestamps, paths)?
  3. Do metric names include units (_seconds, _bytes)?
  4. Are metrics registered once (not per-request)?
  5. Is /metrics endpoint properly exposed?

相关 Skills

前端设计

by anthropics

Universal
热门

面向组件、页面、海报和 Web 应用开发,按鲜明视觉方向生成可直接落地的前端代码与高质感 UI,适合做 landing page、Dashboard 或美化现有界面,避开千篇一律的 AI 审美。

想把页面做得既能上线又有设计感,就用前端设计:组件到整站都能产出,难得的是能避开千篇一律的 AI 味。

编码与调试
未扫描109.6k

网页构建器

by anthropics

Universal
热门

面向复杂 claude.ai HTML artifact 开发,快速初始化 React + Tailwind CSS + shadcn/ui 项目并打包为单文件 HTML,适合需要状态管理、路由或多组件交互的页面。

在 claude.ai 里做复杂网页 Artifact 很省心,多组件、状态和路由都能顺手搭起来,React、Tailwind 与 shadcn/ui 组合效率高、成品也更精致。

编码与调试
未扫描109.6k

网页应用测试

by anthropics

Universal
热门

用 Playwright 为本地 Web 应用编写自动化测试,支持启动开发服务器、校验前端交互、排查 UI 异常、抓取截图与浏览器日志,适合调试动态页面和回归验证。

借助 Playwright 一站式验证本地 Web 应用前端功能,调 UI 时还能同步查看日志和截图,定位问题更快。

编码与调试
未扫描109.6k

相关 MCP 服务

GitHub

编辑精选

by GitHub

热门

GitHub 是 MCP 官方参考服务器,让 Claude 直接读写你的代码仓库和 Issues。

这个参考服务器解决了开发者想让 AI 安全访问 GitHub 数据的问题,适合需要自动化代码审查或 Issue 管理的团队。但注意它只是参考实现,生产环境得自己加固安全。

编码与调试
82.9k

by Context7

热门

Context7 是实时拉取最新文档和代码示例的智能助手,让你告别过时资料。

它能解决开发者查找文档时信息滞后的问题,特别适合快速上手新库或跟进更新。不过,依赖外部源可能导致偶尔的数据延迟,建议结合官方文档使用。

编码与调试
51.5k

by tldraw

热门

tldraw 是让 AI 助手直接在无限画布上绘图和协作的 MCP 服务器。

这解决了 AI 只能输出文本、无法视觉化协作的痛点——想象让 Claude 帮你画流程图或白板讨论。最适合需要快速原型设计或头脑风暴的开发者。不过,目前它只是个基础连接器,你得自己搭建画布应用才能发挥全部潜力。

编码与调试
46.2k

评论