OpenTelemetry实践系列-1 部署服务篇

两行法桐2026-08-03 9:55

环境准备

yaml 复制代码 
## 操作系统：ubuntu 22.04
master:
192.168.10.128
worker:
192.168.10.129
nfs server/harbor:
192.168.10.136

创建nfs和部署nfs-subdir-external-provisioner

192.168.10.136节点部署nfs服务端

yaml 复制代码 
# 在 192.168.10.136 上执行

sudo apt update
sudo apt install -y nfs-kernel-server

sudo mkdir -p /data/nfs
sudo chmod 777 /data/nfs

sudo tee /etc/exports <<EOF
/data/nfs 192.168.10.128(rw,sync,no_subtree_check,no_root_squash)
/data/nfs 192.168.10.129(rw,sync,no_subtree_check,no_root_squash)
EOF

sudo exportfs -ra
sudo systemctl enable nfs-kernel-server
sudo systemctl restart nfs-kernel-server

echo "=== NFS 共享列表 ==="
sudo showmount -e localhost

k8s节点部署nfs客户端(master、worker)

yaml 复制代码 
sudo apt update
sudo apt install -y nfs-common

测试 NFS 连通性（任选一个节点）：

yaml 复制代码 
showmount -e 192.168.10.136

# 测试挂载
sudo mkdir -p /mnt/test
sudo mount -t nfs 192.168.10.136:/data/nfs /mnt/test
df -h | grep nfs
sudo umount /mnt/test

在 K8s 中部署 NFS Provisioner

在 master 上执行

yaml 复制代码 
# 添加仓库
helm repo add nfs-subdir-external-provisioner https://kubernetes-sigs.github.io/nfs-subdir-external-provisioner/
helm repo update nfs-subdir-external-provisioner

# 安装（指定 192.168.10.136为 NFS 服务器）
helm install nfs-subdir-external-provisioner nfs-subdir-external-provisioner/nfs-subdir-external-provisioner \
  --namespace kube-system \
  --set nfs.server=192.168.10.136 \
  --set nfs.path=/data/nfs \
  --set storageClass.name=nfs-client \
  --set storageClass.defaultClass=false

发现拉取镜像失败

yaml 复制代码 
kubectl -n kube-system get pod

更换nfs-subdir-external-provisioner镜像为国内镜像

yaml 复制代码 
kubectl -n kube-system edit deploy nfs-subdir-external-provisioner 
## swr.cn-north-4.myhuaweicloud.com/ddn-k8s/registry.k8s.io/sig-storage/nfs-subdir-external-provisioner:v4.0.2

验证

yaml 复制代码 
 kubectl -n kube-system get deploy nfs-subdir-external-provisioner

将 nfs-client 设为默认 StorageClass

shell 复制代码 
kubectl patch storageclass nfs-client -p '{"metadata":{"annotations":{"storageclass.kubernetes.io/is-default-class":"true"}}}'

namespace

yaml 复制代码 
vim ns.yaml
yaml 复制代码 
apiVersion: v1
kind: Namespace
metadata:
  name:  observability
yaml 复制代码 
kubectl apply -f ns.yaml

minio

minio.yaml

yaml 复制代码 
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: minio-pvc
  namespace: observability
spec:
  storageClassName: nfs-client
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: minio
  name: minio
  namespace: observability
spec:
  selector:
    matchLabels:
      app: minio
  template:
    metadata:
      labels:
        app: minio
    spec:
      containers:
      - name: minio
        image: minio/minio:RELEASE.2025-03-12T18-04-18Z
        imagePullPolicy: IfNotPresent
        command: ["/usr/bin/minio"]
        args: ["server", "/data", "--console-address", ":9090"]
        volumeMounts:
        - mountPath: /data
          name: data
        ports:
        - containerPort: 9090
          name: console
        - containerPort: 9000
          name: api
        resources:
          requests:
            cpu: 100m
            memory: 128Mi
          limits:
            cpu: "1"
            memory: 1Gi
      volumes:
      - name: data
        persistentVolumeClaim:
          claimName: minio-pvc
---
apiVersion: v1
kind: Service
metadata:
  name: minio-service
  namespace: observability
spec:
    type: NodePort 
    selector:     
      app: minio
    ports:
    - name: console
      port: 9090
      protocol: TCP
      targetPort: 9090
      nodePort: 30300
    - name: api
      port: 9000
      protocol: TCP
      targetPort: 9000
      nodePort: 30200

部署

yaml 复制代码 
kubectl apply -f minio.yaml
yaml 复制代码 
kubectl -n observability get pod
kubectl -n observability get svc

验证

http://192.168.10.129:30300/login

访问k8s节点ip:30300，默认用户名密码都是minioadmin

minIO配置

创建loki和tempo的bucket

创建 ak、sk 然后赋予相关bucket的权限

AK：kjq5aB9MN5OVLVX1b1nw

SK：MPJfzGvQDNIOzhh5QL0v3WYY8d0EsTXyE3lZP45I

创建授权

json 复制代码 
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "s3:ListAllMyBuckets",
                "s3:ListBucket",
                "s3:GetBucketLocation",
                "s3:GetObject",
                "s3:PutObject",
                "s3:DeleteObject"
            ],
            "Resource": [
                "arn:aws:s3:::loki/*",
                "arn:aws:s3:::tempo/*"
            ]
        }
    ]
}

prometheus grafana

kube-prometheus-stack-values.yaml

json 复制代码 
vim kube-prometheus-stack-values.yaml
json 复制代码 
# ============================================
# kube-prometheus-stack values.yaml
# 适配 K8s 1.34 测试环境
# 节点：master(37) + worker(103, 98)
# NFS StorageClass: nfs-client
# ============================================

# --------------------------------------------
# 全局配置
# --------------------------------------------
global:
  rbac:
    create: true

# --------------------------------------------
# Prometheus Operator
# --------------------------------------------
prometheusOperator:
  enabled: true
  resources:
    limits:
      cpu: 200m
      memory: 256Mi
    requests:
      cpu: 100m
      memory: 128Mi

# --------------------------------------------
# Prometheus
# --------------------------------------------
prometheus:
  enabled: true
  
  prometheusSpec:
    # 数据保留 7 天（测试环境）
    retention: "7d"
    
    # 持久化存储
    storageSpec:
      volumeClaimTemplate:
        spec:
          storageClassName: nfs-client
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 10Gi
    
    # 资源限制
    resources:
      limits:
        cpu: 1000m
        memory: 2Gi
      requests:
        cpu: 500m
        memory: 512Mi
    
    # 服务发现配置
    serviceMonitorSelectorNilUsesHelmValues: false
    serviceMonitorSelector: {}
    podMonitorSelectorNilUsesHelmValues: false
    podMonitorSelector: {}
    
    # 启用 exemplar 存储（Grafana Exemplars 功能的前提）
    enableFeatures:
      - exemplar-storage
    
    # 启用 remote write receiver（OTel Collector 写入 metrics 所需）
    enableRemoteWriteReceiver: true

  service:
    type: NodePort
    nodePort: 30090
    port: 9090
    targetPort: 9090

# --------------------------------------------
# Alertmanager
# --------------------------------------------
alertmanager:
  enabled: true
  
  config:
    global:
      resolve_timeout: 5m
    route:
      group_by: ['alertname', 'job']
      group_wait: 30s
      group_interval: 5m
      repeat_interval: 12h
      receiver: 'null'
      routes:
        - match:
            severity: critical
          receiver: 'null'
    receivers:
      - name: 'null'
  
  alertmanagerSpec:
    # 持久化存储
    storage:
      volumeClaimTemplate:
        spec:
          storageClassName: nfs-client
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 2Gi
    
    resources:
      limits:
        cpu: 200m
        memory: 256Mi
      requests:
        cpu: 100m
        memory: 128Mi

# --------------------------------------------
# Grafana（核心配置）
# --------------------------------------------
grafana:
  enabled: true
  
  # ====== 关键：禁用 chart 自动创建的默认数据源，避免冲突 ======
  defaultDatasourceEnabled: false
  
  adminUser: admin
  adminPassword: admin123
  
  persistence:
    enabled: true
    type: pvc
    storageClassName: nfs-client
    accessModes: ["ReadWriteOnce"]
    size: 5Gi
  
  service:
    type: NodePort
    nodePort: 30030
    port: 80
    targetPort: 3000
  
  resources:
    limits:
      cpu: 500m
      memory: 1Gi
    requests:
      cpu: 100m
      memory: 256Mi
  
  env:
    GF_PLUGINS_PREINSTALL_DISABLED: "true"
  
  # ========== 显式声明 Prometheus 数据源，解锁 Exemplars 配置 ==========
  datasources:
    datasources.yaml:
      apiVersion: 1
      datasources:
        - name: Prometheus
          type: prometheus
          access: proxy
          url: http://kube-prometheus-stack-prometheus.observability.svc.cluster.local:9090
          isDefault: true
          editable: true
          jsonData:
            httpMethod: POST
            # 关键：设置 Prometheus type，否则 Exemplars 区域不会显示 Add 按钮
            prometheusType: Prometheus
            cacheLevel: Low
            manageAlerts: true
            # exemplar 配置留空，由你在 UI 手动配置
            exemplarTraceIdDestinations: []
  
  # ========== 启用 Dashboard 自动导入 ==========
  sidecar:
    dashboards:
      enabled: true
      label: grafana_dashboard
      labelValue: "1"
      # 搜索所有命名空间中的 ConfigMap
      searchNamespace: ALL
      # 更新频率
      reloadOnChange: true
    datasources:
      enabled: true
      label: grafana_datasource
      labelValue: "1"
      searchNamespace: ALL
  
  # 启用默认的 K8s 监控 Dashboard
  defaultDashboardsEnabled: true
  
  # 测试环境不需要高可用
  replicas: 1
  affinity: {}
  
  tolerations:
    - key: node-role.kubernetes.io/control-plane
      operator: Exists
      effect: NoSchedule

# --------------------------------------------
# Node Exporter（节点监控）
# --------------------------------------------
nodeExporter:
  enabled: true

# --------------------------------------------
# kube-state-metrics（K8s 资源监控）
# --------------------------------------------
kube-state-metrics:
  enabled: true
  
  image:
    registry: swr.cn-north-4.myhuaweicloud.com/ddn-k8s
    repository: registry.k8s.io/kube-state-metrics/kube-state-metrics
    tag: v2.19.1
    pullPolicy: IfNotPresent
  
  resources:
    limits:
      cpu: 100m
      memory: 128Mi
    requests:
      cpu: 50m
      memory: 64Mi

# --------------------------------------------
# Prometheus Adapter（HPA 指标，可选）
# --------------------------------------------
prometheusAdapter:
  enabled: false

# --------------------------------------------
# Thanos Sidecar（可选，测试环境关闭）
# --------------------------------------------
thanosSidecar:
  enabled: false

部署命令

json 复制代码 
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts \
  --timeout 300s
helm repo update prometheus-community --timeout 300s
helm install kube-prometheus-stack prometheus-community/kube-prometheus-stack \
  --namespace observability \
  -f kube-prometheus-stack-values.yaml \
  --timeout 10m

## 如果上面因为网络报错，可使用下面的wget命令直接下载，或者通过链接离线下载并上传
wget https://github.com/prometheus-community/helm-charts/releases/download/kube-prometheus-stack-87.1.0/kube-prometheus-stack-87.1.0.tgz
helm install kube-prometheus-stack ./kube-prometheus-stack-87.1.0.tgz \
  --namespace observability \
  -f kube-prometheus-stack-values.yaml \
  --timeout 10m
json 复制代码 
# 查看所有组件状态
kubectl -n observability get pod

验证

访问 Grafana

http://任意节点IP:30030

账号：admin / admin123

查看 Prometheus

json 复制代码 
kubectl -n observability port-forward svc/kube-prometheus-stack-prometheus 9090:9090

浏览器访问 http://任意节点IP:30090/

卸载

json 复制代码 
kubectl -n observability get all
helm uninstall kube-prometheus-stack -n observability
kubectl -n observability delete pvc kube-prometheus-stack-grafana

tempo

tempo-values.yaml

注意替换成自己的ak和sk

yaml 复制代码 
vim tempo-values.yaml
yaml 复制代码 
# ============================================
# Tempo Distributed - 测试环境配置
# 数据存储：MinIO 
# ============================================

# 启用 HTTP 流式查询（Grafana 搜索trace需要）
streamOverHTTPEnabled: true

# --------------------------------------------
# 1. 存储配置：使用外部 MinIO (S3 兼容)
# --------------------------------------------
storage:
  trace:
    backend: s3
    s3:
      bucket: tempo
      endpoint: minio-service.observability.svc.cluster.local:9000
      region: cn-north-1
      access_key: kjq5aB9MN5OVLVX1b1nw
      secret_key: MPJfzGvQDNIOzhh5QL0v3WYY8d0EsTXyE3lZP45I
      insecure: true

# --------------------------------------------
# 2. 禁用内置 MinIO
# --------------------------------------------
minio:
  enabled: false

# --------------------------------------------
# 3. 接收器配置
# --------------------------------------------
traces:
  otlp:
    grpc:
      enabled: true
      port: 4317
    http:
      enabled: true
      port: 4318
  zipkin:
    enabled: false
  jaeger:
    thriftCompact:
      enabled: false
    thriftHttp:
      enabled: false
  opencensus:
    enabled: false

# --------------------------------------------
# 4. Ingester：3 副本，测试环境允许共节点
# --------------------------------------------
ingester:
  replicas: 3
  persistence:
    enabled: false
  # 关键修复：affinity 必须是完整对象，不能用数组语法覆盖
  affinity: {}
  resources:
    requests:
      cpu: 100m
      memory: 256Mi
    limits:
      cpu: 500m
      memory: 512Mi

# --------------------------------------------
# 5. Distributor（合并后的唯一配置）
# --------------------------------------------
distributor:
  replicas: 1
  config:
    log_received_spans:
      enabled: true
  resources:
    requests:
      cpu: 50m
      memory: 128Mi
    limits:
      cpu: 200m
      memory: 256Mi

# --------------------------------------------
# 6. 其他组件
# --------------------------------------------
querier:
  replicas: 1
  resources:
    requests:
      cpu: 50m
      memory: 128Mi
    limits:
      cpu: 200m
      memory: 256Mi

queryFrontend:
  replicas: 1
  resources:
    requests:
      cpu: 50m
      memory: 128Mi
    limits:
      cpu: 200m
      memory: 256Mi

compactor:
  replicas: 1
  resources:
    requests:
      cpu: 100m
      memory: 512Mi    # 从 128Mi 提升到 512Mi
    limits:
      cpu: 500m
      memory: 1Gi      # 从 256Mi 提升到 1Gi      

# --------------------------------------------
# 7. Metrics Generator
# --------------------------------------------
metricsGenerator:
  enabled: true
  replicas: 1
  resources:
    requests:
      cpu: 50m
      memory: 128Mi
    limits:
      cpu: 200m
      memory: 256Mi

# --------------------------------------------
# 8. Memcached
# --------------------------------------------
memcached:
  enabled: true
  replicas: 1
  resources:
    requests:
      cpu: 50m
      memory: 128Mi
    limits:
      cpu: 100m
      memory: 256Mi

# --------------------------------------------
# 9. Gateway
# --------------------------------------------
gateway:
  enabled: true
  replicas: 1
  resources:
    requests:
      cpu: 50m
      memory: 64Mi
    limits:
      cpu: 100m
      memory: 128Mi

部署

bash 复制代码 
helm repo add grafana https://grafana.github.io/helm-charts 
helm repo update grafana
## 网络原因会需要尝试多次
helm install tempo grafana/tempo-distributed \
  --namespace observability \
  -f tempo-values.yaml \
  --timeout 10m \
  --wait


##如果部署失败，发现pod的日志提示amd不支持，则可以部署下面的
helm install tempo grafana/tempo-distributed \
  -n observability \
  --version 1.8.0 \
  --set global.image.tag=2.3.1 \
  -f tempo-values.yaml --timeout 10m
bash 复制代码 
kubectl -n observability get pod

验证

添加 Tempo 数据源

  • URL : [http://tempo-gateway.observability.svc.cluster.local](http://tempo-gateway.observability.svc.cluster.local)
  • 类型: Tempo

卸载

bash 复制代码 
helm uninstall tempo -n observability

loki

loki-values.yaml

注意替换成自己的ak和sk

yaml 复制代码 
vim loki-values.yaml
yaml 复制代码 
# ============================================
# Loki Helm Chart 自定义配置 (Simple Scalable Mode)
# 适配 K8s 1.34 测试环境
# 存储后端：外部 MinIO (S3 兼容)
# ============================================

# 部署模式：SimpleScalable（适合测试环境）
deploymentMode: SimpleScalable

# 禁用内置 MinIO（使用外部 MinIO）
minio:
  enabled: false

# 关键：忽略 MinIO 弃用警告（chart v17+ 需要）
ignoreMinioDeprecation: true

# Loki 核心配置
loki:
  # Schema 配置
  schemaConfig:
    configs:
      - from: "2024-04-01"
        store: tsdb
        object_store: s3
        schema: v13
        index:
          prefix: loki_index_
          period: 24h

  # 存储配置：使用外部 MinIO 的 S3 接口
  storage:
    type: s3
    bucketNames:
      ## 使用minio的bucket    
      chunks: loki
      ruler: loki
      admin: loki
    s3:
      endpoint: minio-service.observability.svc.cluster.local:9000
      region: cn-north-1
      accessKeyId: "kjq5aB9MN5OVLVX1b1nw"
      secretAccessKey: "MPJfzGvQDNIOzhh5QL0v3WYY8d0EsTXyE3lZP45I"      
      insecure: true
      s3forcepathstyle: true

  # 压缩配置
  ingester:
    lifecycler:
      ring:
        replication_factor: 1  # 测试环境改为 1  
    chunk_encoding: snappy

  # 查询配置
  querier:
    max_concurrent: 4

  # 模式分析器
  pattern_ingester:
    enabled: true

  # 限制配置
  limits_config:
    allow_structured_metadata: true
    volume_enabled: true
    retention_period: 168h  # 7天保留

  # 通用配置
  auth_enabled: false

# ============================================
# 组件配置
# ============================================

# backend 需要持久化（存储索引和规则）
backend:
  replicas: 1
  persistence:
    enabled: true
    storageClass: nfs-client
    size: 5Gi
  resources:
    limits:
      cpu: 500m
      memory: 1Gi
    requests:
      cpu: 100m
      memory: 256Mi

# read 是无状态组件，不需要持久化存储
read:
  replicas: 1
  resources:
    limits:
      cpu: 500m
      memory: 1Gi
    requests:
      cpu: 100m
      memory: 256Mi

# write 需要持久化（缓存日志块）
write:
  replicas: 1
  persistence:
    enabled: true
    storageClass: nfs-client
    size: 10Gi
  resources:
    limits:
      cpu: 500m
      memory: 1Gi
    requests:
      cpu: 100m
      memory: 256Mi

# Gateway 配置（NGINX 入口）
gateway:
  enabled: true
  service:
    type: NodePort
    nodePort: 30032
  # 禁用 access-log-exporter（AMD64 v2 不兼容）
  accessLogs:
    enabled: false
  resources:
    limits:
      cpu: 200m
      memory: 256Mi
    requests:
      cpu: 50m
      memory: 64Mi

# 缓存配置（测试环境可以关闭或简化）
chunksCache:
  enabled: false

resultsCache:
  enabled: false

# Loki Canary（日志监控探针）
lokiCanary:
  enabled: true

# 容忍调度到 master
tolerations:
  - key: node-role.kubernetes.io/control-plane
    operator: Exists
    effect: NoSchedule

部署

shell 复制代码 
helm repo add grafana-community https://grafana-community.github.io/helm-charts 
helm update grafana-community

helm install loki grafana-community/loki \
  --namespace observability \
  -f loki-values.yaml \
  --timeout 10m \
  --wait

##网络不行的话，就手动下载
wget https://github.com/grafana-community/helm-charts/releases/download/loki-18.1.1/loki-18.1.1.tgz
helm upgrade --install loki ./loki-18.1.1.tgz \
  --namespace observability \
  -f loki-values.yaml \
  --timeout 10m \
  --wait

## 如果exporter有问题则去掉exporter
kubectl patch deployment loki-gateway -n observability --type='json' -p='[
  {
    "op": "remove",
    "path": "/spec/template/spec/containers/1"
  }
]'
json 复制代码 
# 查看所有组件状态
kubectl -n observability get pod

验证

添加数据源

shell 复制代码 
http://loki-gateway.observability.svc.cluster.local:80
shell 复制代码 
NODE_IP=$(kubectl get nodes -o jsonpath='{.items[0].status.addresses[0].address}')

# 推送测试日志
curl -X POST "http://${NODE_IP}:30032/loki/api/v1/push" \
  -H "Content-Type: application/json" \
  -d '{
    "streams": [{
      "stream": {"job":"test","level":"info","app":"myapp"},
      "values": [
        ["'$(date +%s%N)'", "测试日志 1: Hello Loki"],
        ["'$(($(date +%s%N)+1000000000))'", "测试日志 2: 这是一条测试"],
        ["'$(($(date +%s%N)+2000000000))'", "测试日志 3: Grafana 可以查询了"]
      ]
    }]
  }'

echo "推送完成，等待 2 秒..."
sleep 2
# 使用 range query 查询（最近 5 分钟）
START=$(date -d '5 minutes ago' +%s)000000000
END=$(date +%s)000000000

# 验证查询
curl -s "http://${NODE_IP}:30032/loki/api/v1/query_range?query=%7Bjob%3D%22test%22%7D&start=${START}&end=${END}&limit=100" | python3 -m json.tool 2>/dev/null || cat

推送后，在 Grafana Explore 中输入：

shell 复制代码 
{job="test"}

点击 Run query，应该能看到 3 条日志。

卸载

bash 复制代码 
kubectl delete pvc -n observability -l app.kubernetes.io/name=loki
helm uninstall loki -n observability

promtail

promtail-values.yaml

bash 复制代码 
config:
  clients:
    - url: http://loki-gateway.observability.svc.cluster.local/loki/api/v1/push

  snippets:
    extraScrapeConfigs: |
      - job_name: kubernetes-pods-custom
        kubernetes_sd_configs:
          - role: pod
        pipeline_stages:
          - cri: {}
        relabel_configs:
          - source_labels: [__meta_kubernetes_namespace]
            target_label: namespace
          - source_labels: [__meta_kubernetes_pod_name]
            target_label: pod
          - source_labels: [__meta_kubernetes_pod_container_name]
            target_label: container
          - action: drop
            source_labels: [__meta_kubernetes_pod_phase]
            regex: Succeeded|Failed|Completed

resources:
  limits:
    cpu: 500m
    memory: 512Mi
  requests:
    cpu: 100m
    memory: 128Mi

tolerations:
  - key: node-role.kubernetes.io/control-plane
    operator: Exists
    effect: NoSchedule

部署

bash 复制代码 
helm repo add grafana https://grafana.github.io/helm-charts
helm repo update grafana

helm upgrade --install promtail grafana/promtail \
  -n observability \
  -f promtail-values.yaml \
  --timeout 10m \
  --wait
json 复制代码 
# 查看所有组件状态
kubectl -n observability get pod

卸载

bash 复制代码 
helm uninstall promtail -n observability

在loki的数据源设置租户id X-Scope-OrgID =1

部署 OpenTelemetry Collector

otel-collector.yaml

bash 复制代码 
vim  otel-collector.yaml
yaml 复制代码 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: otel-collector
  namespace: observability
spec:
  replicas: 1
  selector:
    matchLabels:
      app: otel-collector
  template:
    metadata:
      labels:
        app: otel-collector
    spec:
      tolerations:
        - key: node-role.kubernetes.io/control-plane
          operator: Exists
          effect: NoSchedule
      containers:
        - name: otel-collector
          image: otel/opentelemetry-collector-contrib:0.104.0
          args:
            - --config=/etc/otel-collector-config/otel-collector-config.yaml
          ports:
            - containerPort: 4317
              name: otlp-grpc
            - containerPort: 4318
              name: otlp-http
            - containerPort: 13133
              name: health
          volumeMounts:
            - name: config
              mountPath: /etc/otel-collector-config
      volumes:
        - name: config
          configMap:
            name: otel-collector-config
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: otel-collector-config
  namespace: observability
data:
  otel-collector-config.yaml: |
    extensions:
      health_check:
        endpoint: 0.0.0.0:13133
    
    receivers:
      otlp:
        protocols:
          grpc:
            endpoint: 0.0.0.0:4317
          http:
            endpoint: 0.0.0.0:4318
    
    processors:
      batch:
        timeout: 1s
        send_batch_size: 1024
      resource:
        attributes:
          - key: service.namespace
            value: "default"
            action: upsert
    
    exporters:
      otlp/tempo:
        endpoint: tempo-distributor.observability.svc.cluster.local:4317
        tls:
          insecure: true
      
      # loki exporter 0.104.0 最简配置
      loki:
        endpoint: http://loki-gateway.observability.svc.cluster.local:80/loki/api/v1/push
        tls:
          insecure: true
      
      prometheusremotewrite:
        endpoint: http://kube-prometheus-stack-prometheus.observability.svc.cluster.local:9090/api/v1/write
        tls:
          insecure: true
        external_labels:
          collector: otel-collector
    
    service:
      extensions: [health_check]
      telemetry:
        metrics:
          address: 0.0.0.0:8889
      pipelines:
        traces:
          receivers: [otlp]
          processors: [batch]
          exporters: [otlp/tempo]
        logs:
          receivers: [otlp]
          processors: [batch, resource]
          exporters: [loki]
        metrics:
          receivers: [otlp]
          processors: [batch, resource]
          exporters: [prometheusremotewrite]
---
apiVersion: v1
kind: Service
metadata:
  name: otel-collector
  namespace: observability
spec:
  selector:
    app: otel-collector
  ports:
    - name: otlp-grpc
      port: 4317
      targetPort: 4317
    - name: otlp-http
      port: 4318
      targetPort: 4318

部署

yaml 复制代码 
kubectl apply -f ./otel-collector.yaml
json 复制代码 
# 查看所有组件状态
kubectl -n observability get pod

卸载

bash 复制代码 
kubectl delete -f ./otel-collector.yaml

OpenTelemetry实践系列-1 部署服务篇环境准备 - 掘金

OpenTelemetry实践系列-2 Log2Trace和Trace2Log - 掘金

OpenTelemetry实践系列-3 Metrics2Trace和Trace2Metrics - 掘金

相关推荐
盛世宏博智慧档案5 小时前
户外配电柜为什么要装POE供电以太网温湿度传感器？
服务器·网络·人工智能·监控·温湿度·配电柜
行百里er1 天前
Spring Insight 里收到的 Span 是怎么存下来的
spring boot·spring·监控
行百里er2 天前
Spring Insight 里上报 Span 为什么用 JDK HttpClient 而不是 RestTemplate
spring boot·spring cloud·监控
SkyWalking中文站3 天前
Horizon UI 1.0 正式发布：SkyWalking 新一代控制台接棒 Booster
运维·监控·自动化运维
行百里er3 天前
BlockingQueue：异步批量上报
spring boot·spring cloud·监控
IT界的老黄牛3 天前
Prometheus TSDB 拆不出来、也存不了一年：4 条外部存储出路 + 容量测算
prometheus·时序数据库·监控·thanos·victoriametrics·remote_write
IT界的老黄牛5 天前
Jenkins 监控一条龙：接入、看板 9964、11 条告警规则直接抄走
jenkins·grafana·prometheus·监控·ci-cd·告警规则
SkyWalking中文站5 天前
SkyWalking 11 与 BanyanDB 0.11：在存储引擎内部实现 Trace 尾部采样
运维·监控·自动化运维
行百里er5 天前
BeanPostProcessor：包装 Feign Client
spring boot·后端·监控
行百里er6 天前
HandlerInterceptor 和 WebFilter：两套 HTTP 请求拦截
后端·架构·监控
热门推荐
01国内可直接用、免费额度/永久免费的大模型API清单（含 SiliconFlow、火山、阿里、智谱、百度、Kimi、DeepSeek、DMXAPI 等）02蜘蛛磁力 搜索引擎大全，如何使用蜘蛛磁力查找磁力链接03GitHub 镜像站点04历年考研数学一、数学二、数学三真题试卷及答案PDF05【一等奖版】2026 国赛 A题 药材的烘干问题06微信历史版本含下载地址（ Windows PC | 安卓 | MAC ）及设置微信不更新072026 国产 AI 大模型横评：DeepSeek、通义千问、Kimi、文心一言、星火、豆包谁更能打？08人工智能（AI）发展史：详细里程碑09ZCode 周末送额度活动开启：3 亿 Token 免费领取10如何新建文件夹？ 电脑新建文件夹的4种方法