Jaeger 分散式追蹤系統

專案簡介

Jaeger 是由 Uber 開發的開源分散式追蹤系統，現為 CNCF 畢業專案。用於監控和排查微服務架構中的問題，支援 OpenTelemetry。

GitHub Stars: 22K+

主要功能

分散式追蹤 - 端對端請求追蹤
根因分析 - 定位效能問題
服務相依 - 自動產生服務圖
效能優化 - 識別延遲瓶頸
OpenTelemetry - 原生整合

架構元件

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│   Agent     │───▶│  Collector  │───▶│   Storage   │
└─────────────┘    └─────────────┘    └─────────────┘
                          │
                          ▼
                   ┌─────────────┐
                   │    Query    │
                   └─────────────┘
                          │
                          ▼
                   ┌─────────────┐
                   │     UI      │
                   └─────────────┘

安裝

All-in-One（開發用）

1
2
3
4
5
6
7
docker run -d --name jaeger \
  -p 16686:16686 \
  -p 4317:4317 \
  -p 4318:4318 \
  -p 14250:14250 \
  -p 14268:14268 \
  jaegertracing/all-in-one:latest

訪問 http://localhost:16686

Docker Compose

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
version: '3.8'
services:
  jaeger:
    image: jaegertracing/all-in-one:latest
    ports:
      - "16686:16686"   # UI
      - "4317:4317"     # OTLP gRPC
      - "4318:4318"     # OTLP HTTP
      - "14268:14268"   # Thrift HTTP
    environment:
      - COLLECTOR_OTLP_ENABLED=true

Kubernetes

1
2
kubectl create namespace observability
kubectl apply -f https://github.com/jaegertracing/jaeger-operator/releases/download/v1.50.0/jaeger-operator.yaml -n observability

OpenTelemetry 整合

Python

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
from opentelemetry import trace
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.sdk.resources import Resource

resource = Resource.create({"service.name": "my-service"})

provider = TracerProvider(resource=resource)
processor = BatchSpanProcessor(OTLPSpanExporter(endpoint="http://localhost:4317"))
provider.add_span_processor(processor)
trace.set_tracer_provider(provider)

tracer = trace.get_tracer(__name__)

# 使用
with tracer.start_as_current_span("operation") as span:
    span.set_attribute("user.id", "12345")
    # 業務邏輯

Go

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
package main

import (
    "go.opentelemetry.io/otel"
    "go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracegrpc"
    "go.opentelemetry.io/otel/sdk/trace"
)

func initTracer() func() {
    exporter, _ := otlptracegrpc.New(context.Background(),
        otlptracegrpc.WithEndpoint("localhost:4317"),
        otlptracegrpc.WithInsecure(),
    )

    tp := trace.NewTracerProvider(
        trace.WithBatcher(exporter),
        trace.WithResource(resource.NewWithAttributes(
            semconv.ServiceNameKey.String("my-service"),
        )),
    )

    otel.SetTracerProvider(tp)
    return func() { tp.Shutdown(context.Background()) }
}

Node.js

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
const { NodeSDK } = require("@opentelemetry/sdk-node");
const {
  OTLPTraceExporter,
} = require("@opentelemetry/exporter-trace-otlp-grpc");

const sdk = new NodeSDK({
  traceExporter: new OTLPTraceExporter({
    url: "http://localhost:4317",
  }),
  serviceName: "my-node-service",
});

sdk.start();

儲存後端

Elasticsearch

1
2
3
4
5
6
7
8
services:
  jaeger-collector:
    environment:
      - SPAN_STORAGE_TYPE=elasticsearch
      - ES_SERVER_URLS=http://elasticsearch:9200

  elasticsearch:
    image: elasticsearch:8.x

Cassandra

1
2
3
4
5
6
7
8
services:
  jaeger-collector:
    environment:
      - SPAN_STORAGE_TYPE=cassandra
      - CASSANDRA_SERVERS=cassandra

  cassandra:
    image: cassandra:4.x

Kafka（緩衝）

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
services:
  jaeger-collector:
    environment:
      - SPAN_STORAGE_TYPE=kafka
      - KAFKA_PRODUCER_BROKERS=kafka:9092

  jaeger-ingester:
    environment:
      - SPAN_STORAGE_TYPE=elasticsearch
      - KAFKA_CONSUMER_BROKERS=kafka:9092

查詢和分析

搜尋 Trace

1
2
3
4
5
6
# UI 查詢
Service: my-service
Operation: /api/users
Tags: http.status_code=500
Min Duration: 1s
Limit: 20

Trace 比較

選擇多個 Trace
點擊 Compare
分析差異

服務效能

P50/P95/P99 延遲
請求率
錯誤率
服務相依圖

Span 屬性

設定屬性

1
2
3
4
5
6
7
8
from opentelemetry import trace

tracer = trace.get_tracer(__name__)

with tracer.start_as_current_span("db-query") as span:
    span.set_attribute("db.system", "postgresql")
    span.set_attribute("db.statement", "SELECT * FROM users")
    span.set_attribute("db.name", "mydb")

標準屬性

屬性	說明
`http.method`	HTTP 方法
`http.url`	請求 URL
`http.status_code`	狀態碼
`db.system`	資料庫類型
`db.statement`	SQL 語句
`rpc.system`	RPC 系統

取樣策略

Collector 設定

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
# sampling.json
{
  "service_strategies": [
    {
      "service": "my-service",
      "type": "probabilistic",
      "param": 0.5
    }
  ],
  "default_strategy": {
    "type": "probabilistic",
    "param": 0.1
  }
}

取樣類型

類型	說明
`const`	固定取樣（0 或 1）
`probabilistic`	機率取樣
`ratelimiting`	速率限制
`remote`	遠端控制

Jaeger Operator

部署

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
apiVersion: jaegertracing.io/v1
kind: Jaeger
metadata:
  name: production
spec:
  strategy: production
  storage:
    type: elasticsearch
    options:
      es:
        server-urls: http://elasticsearch:9200
  collector:
    replicas: 2
  query:
    replicas: 2

自動注入

1
2
3
4
5
apiVersion: apps/v1
kind: Deployment
metadata:
  annotations:
    sidecar.jaegertracing.io/inject: "true"

Jaeger 分散式追蹤系統

使用 Jaeger 追蹤微服務架構中的請求流程，定位效能瓶頸和問題根因

專案簡介

主要功能

架構元件

安裝

All-in-One（開發用）

Docker Compose

Kubernetes

OpenTelemetry 整合

Python

Go

Node.js

儲存後端

Elasticsearch

Cassandra

Kafka（緩衝）

查詢和分析

搜尋 Trace

Trace 比較

服務效能

Span 屬性

設定屬性

標準屬性

取樣策略

Collector 設定

取樣類型

Jaeger Operator

部署

自動注入

相關連結

延伸閱讀