傳統的性能監控是發現問題後人工排查。2025 年,AI 不僅能發現問題,還能自動建議甚至自動修復。
監控架構
用户瀏覽器
│
▼
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ 數據採集 │────▶│ AI 分析引擎 │────▶│ 自動修復 │
│ (web-vitals)│ │ (Claude) │ │ (PR/配置) │
└─────────────┘ └─────────────┘ └─────────────┘
│ │ │
▼ ▼ ▼
原始指標 性能報告 + 診斷 自動創建修復 PR
數據採集
ts
// lib/performance-monitor.ts
import { onLCP, onINP, onCLS, onTTFB, type Metric } from "web-vitals";
interface PerformanceEntry extends Metric {
url: string;
userAgent: string;
connectionType: string;
deviceMemory: number;
timestamp: number;
}
function collectVitals() {
const baseData = {
url: window.location.href,
userAgent: navigator.userAgent,
connectionType: (navigator as any).connection?.effectiveType ?? "unknown",
deviceMemory: (navigator as any).deviceMemory ?? 0,
timestamp: Date.now(),
};
const report = (metric: Metric) => {
const entry: PerformanceEntry = { ...metric, ...baseData };
// 上報到分析平台
fetch("/api/vitals", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify(entry),
// 用 sendBeacon 保證頁面關閉時也能發送
keepalive: true,
}).catch(() => {
navigator.sendBeacon?.("/api/vitals", JSON.stringify(entry));
});
// 性能差時立即告警
if (entry.rating === "poor") {
console.warn(`[Performance] ${entry.name} = ${entry.value} (${entry.rating})`);
}
};
onLCP(report);
onINP(report);
onCLS(report);
onTTFB(report);
}
// 路由切換時重新採集(SPA 場景)
let lastPath = window.location.pathname;
new PerformanceObserver(() => {
if (window.location.pathname !== lastPath) {
lastPath = window.location.pathname;
collectVitals();
}
}).observe({ type: "navigation", buffered: true });
AI 分析引擎
ts
// scripts/performance-analyzer.ts
import Anthropic from "@anthropic-ai/sdk";
const client = new Anthropic();
interface PerfReport {
page: string;
metrics: {
lcp: { p50: number; p75: number; p95: number };
inp: { p50: number; p75: number; p95: number };
cls: { p50: number; p75: number; p95: number };
ttfb: { p50: number; p75: number; p95: number };
};
trend: "improving" | "stable" | "degrading";
regressions: string[];
}
async function analyzePerformance(report: PerfReport) {
const response = await client.messages.create({
model: "claude-sonnet-4-20250514",
max_tokens: 4096,
system: `你是一個前端性能專家。分析性能數據,給出具體可執行的優化建議。
輸出 JSON 格式:
{
"diagnosis": "問題診斷",
"rootCauses": ["原因1", "原因2"],
"suggestions": [
{
"priority": "high|medium|low",
"action": "具體操作",
"estimatedImpact": "預期提升",
"codeExample": "代碼示例(可選)"
}
]
}`,
messages: [
{
role: "user",
content: `分析以下頁面性能數據:
頁面:${report.page}
趨勢:${report.trend}
迴歸:${report.regressions.join(", ")}
LCP: P50=${report.metrics.lcp.p50}ms, P75=${report.metrics.lcp.p75}ms, P95=${report.metrics.lcp.p95}ms
INP: P50=${report.metrics.inp.p50}ms, P75=${report.metrics.inp.p75}ms, P95=${report.metrics.inp.p95}ms
CLS: P50=${report.metrics.cls.p50}, P75=${report.metrics.cls.p75}, P95=${report.metrics.cls.p95}
TTFB: P50=${report.metrics.ttfb.p50}ms, P75=${report.metrics.ttfb.p75}ms, P95=${report.metrics.ttfb.p95}ms
Core Web Vitals 閾值:LCP<2.5s, INP<200ms, CLS<0.1`,
},
],
});
return JSON.parse(response.content[0].type === "text" ? response.content[0].text : "{}");
}
自動優化建議
tsx
// AI 生成的實際優化建議示例
// 診斷:LCP P75 = 4.2s,超過閾值
// 原因:首屏大圖未 preload,字體阻塞渲染
// 建議 1:Preload 關鍵資源
// app/layout.tsx
export default function RootLayout({ children }: { children: React.ReactNode }) {
return (
<html>
<head>
{/* preload LCP 圖片 */}
<link
rel="preload"
as="image"
href="/hero-banner.webp"
fetchPriority="high"
/>
{/* preload 關鍵字體 */}
<link
rel="preload"
as="font"
href="/fonts/inter-var.woff2"
type="font/woff2"
crossOrigin="anonymous"
/>
</head>
<body>{children}</body>
</html>
);
}
// 建議 2:圖片優化
// components/HeroBanner.tsx
import Image from "next/image";
export function HeroBanner() {
return (
<Image
src="/hero-banner.webp"
width={1920}
height={800}
priority // 自動 preload + fetchPriority="high"
sizes="100vw"
alt="首屏"
// 自動生成 srcset,按需加載不同尺寸
/>
);
}
// 建議 3:字體優化
// 用 font-display: swap 避免字體阻塞
@font-face {
font-family: "Inter";
src: url("/fonts/inter-var.woff2") format("woff2");
font-display: swap; // 先用系統字體,字體加載後替換
}
迴歸檢測
ts
// 性能迴歸自動告警
function checkRegression(current: PerfReport, baseline: PerfReport) {
const regressions: string[] = [];
const THRESHOLD = 0.2; // 20% 以上算迴歸
for (const metric of ["lcp", "inp", "cls", "ttfb"] as const) {
const currentP75 = current.metrics[metric].p75;
const baselineP75 = baseline.metrics[metric].p75;
const change = (currentP75 - baselineP75) / baselineP75;
if (change > THRESHOLD) {
regressions.push(
`${metric.toUpperCase()} P75 從 ${baselineP75} 升至 ${currentP75}(+${(change * 100).toFixed(1)}%)`,
);
}
}
if (regressions.length > 0) {
// 自動創建 Issue 或發送告警
notifyTeam({
title: "性能迴歸告警",
body: regressions.join("\n"),
severity: regressions.length > 2 ? "critical" : "warning",
});
}
}
小結
- 性能監控要採集真實用户數據(RUM),不能只依賴實驗室數據
- AI 不僅能分析性能數據,還能給出具體可執行的優化建議
- 自動迴歸檢測能在線上性能變差時及時告警
- 性能優化是持續的過程,不是一次性的任務
- 把性能指標加入 CI 門禁,防止迴歸