当前位置：首页 > news >正文

全链路追踪：OpenTelemetry与Jaeger实战

news 2026/6/27 7:26:20

全链路追踪：OpenTelemetry与Jaeger实战

大家好，我是欧阳瑞（Rich Own）。今天想和大家聊聊全链路追踪这个重要话题。作为一个全栈开发者，在微服务架构中，全链路追踪是定位问题和性能优化的关键工具。今天就来分享一下OpenTelemetry和Jaeger的实战经验。

全链路追踪概述

为什么需要全链路追踪？

问题	说明
分布式系统复杂	调用链路过长难以追踪
性能瓶颈难定位	无法确定哪个服务慢
故障排查困难	错误传播难以追踪
用户体验监控	需要端到端延迟分析

核心概念

概念	说明
Trace	一次完整的请求链路
Span	链路中的单个操作
Span Context	跨服务传递的上下文
Baggage	用户自定义数据

OpenTelemetry入门

安装依赖

npm install @opentelemetry/sdk-node @opentelemetry/api npm install @opentelemetry/exporter-jaeger @opentelemetry/auto-instrumentations-node

基本配置

// tracing.js const { NodeSDK } = require('@opentelemetry/sdk-node'); const { ConsoleSpanExporter } = require('@opentelemetry/sdk-trace-base'); const { JaegerExporter } = require('@opentelemetry/exporter-jaeger'); const { getNodeAutoInstrumentations } = require('@opentelemetry/auto-instrumentations-node'); const sdk = new NodeSDK({ traceExporter: new JaegerExporter({ endpoint: 'http://localhost:14268/api/traces' }), instrumentations: [getNodeAutoInstrumentations()] }); sdk.start();

手动创建Span

const { trace, context } = require('@opentelemetry/api'); const tracer = trace.getTracer('my-service'); async function processRequest(req) { return tracer.startActiveSpan('process-request', async (span) => { span.setAttribute('request.id', req.id); try { const user = await fetchUser(req.userId); const order = await createOrder(user); return order; } finally { span.end(); } }); }

Jaeger配置

启动Jaeger

# docker-compose.yml version: '3.7' services: jaeger: image: jaegertracing/all-in-one:latest ports: - "5775:5775/udp" - "6831:6831/udp" - "6832:6832/udp" - "5778:5778" - "16686:16686" - "14268:14268" - "9411:9411" environment: - COLLECTOR_ZIPKIN_HOST_PORT=:9411

访问Jaeger UI

# 启动服务 docker-compose up -d # 访问仪表盘 open http://localhost:16686

实战案例：分布式追踪

// 服务A const tracer = trace.getTracer('service-a'); async function handleRequest(req) { return tracer.startActiveSpan('handle-request', async (span) => { span.setAttribute('service', 'service-a'); // 调用服务B const response = await fetch('http://service-b/api/data', { headers: { 'traceparent': trace.getSpan(context.active()).spanContext().traceId } }); span.end(); return response.json(); }); } // 服务B async function handleRequest(req) { return tracer.startActiveSpan('process-data', async (span) => { span.setAttribute('service', 'service-b'); // 调用数据库 const data = await db.query('SELECT * FROM users'); span.end(); return data; }); }

最佳实践

1. 自定义Instrumentation

const { InstrumentationBase } = require('@opentelemetry/instrumentation'); class MyInstrumentation extends InstrumentationBase { constructor() { super('my-instrumentation', '1.0.0'); } init() { // 自定义instrumentation逻辑 } }

2. 采样策略

const { ParentBasedSampler, AlwaysOnSampler, TraceIdRatioBasedSampler } = require('@opentelemetry/sdk-trace-base'); const sampler = new ParentBasedSampler({ root: new TraceIdRatioBasedSampler(0.1) // 10%采样率 });

3. 指标集成

const { MeterProvider } = require('@opentelemetry/sdk-metrics'); const meterProvider = new MeterProvider(); const meter = meterProvider.getMeter('my-service'); const requestCounter = meter.createCounter('requests', { description: 'Total requests' }); // 使用计数器 requestCounter.add(1, { status: 'success' });