修复业务侧日志采集失败问题

deployments/observability/README.md

@@ -0,0 +1,174 @@
+# chat-deploy 可观测性部署指南
+
+本目录包含将 chat-deploy 接入 itom-platform 可观测中心所需的配置文件。
+
+## 功能说明
+
+- **日志采集**：通过 Promtail 采集 Docker 容器日志，推送到 itom-platform 的 Loki
+- **MongoDB 监控**：通过 MongoDB Exporter 采集 MongoDB 指标
+- **Redis 监控**（可选）：通过 Redis Exporter 采集 Redis 指标
+- **系统监控**（可选）：通过 Node Exporter 采集系统级指标（CPU/Memory/Disk/Network）
+
+## 快速开始
+
+### 1. 配置环境变量
+
+复制示例配置文件并修改：
+
+```bash
+cp config.env.example config.env
+```
+
+**必须修改的配置项**：
+
+| 配置项 | 说明 | 示例 |
+|--------|------|------|
+| `OBS_HOST` | itom-platform 可观测中心地址 | `192.168.1.100` 或 `obs.example.com` |
+| `OBS_AUTH_TOKEN` | 鉴权 token（与中心侧一致） | `your-secret-token` |
+| `MONGODB_URI` | MongoDB 连接 URI | `mongodb://user:pass@host:27017/db` |
+
+### 2. 启动日志采集
+
+仅启用日志采集：
+
+```bash
+docker compose --env-file config.env -f docker-compose-observability.yaml up -d
+```
+
+### 3. 启动完整可观测性（日志 + 指标）
+
+同时启用日志和指标采集：
+
+```bash
+docker compose --env-file config.env -f docker-compose-observability.yaml --profile metrics up -d
+```
+
+### 4. 验证服务状态
+
+```bash
+# 查看服务状态
+docker compose --env-file config.env -f docker-compose-observability.yaml ps
+
+# 查看 Promtail 日志
+docker logs chat-deploy-promtail
+
+# 查看 MongoDB Exporter 指标
+curl http://localhost:9216/metrics | head -50
+```
+
+## 配置说明
+
+### MongoDB 配置
+
+MongoDB Exporter 需要正确的连接 URI。根据 `config/mongodb.yml` 中的配置：
+
+```env
+# 默认配置
+MONGODB_URI=mongodb://openIM:openIM123@localhost:37017/openim_v3?authSource=openim_v3
+```
+
+如果 MongoDB 运行在 Docker 容器中，请使用容器名或宿主机 IP：
+
+```env
+# 使用容器名（需要在同一网络）
+MONGODB_URI=mongodb://openIM:openIM123@mongo:27017/openim_v3?authSource=openim_v3
+
+# 使用宿主机 IP
+MONGODB_URI=mongodb://openIM:openIM123@host.docker.internal:37017/openim_v3?authSource=openim_v3
+```
+
+### Redis 配置（可选）
+
+如果需要监控 Redis：
+
+```env
+REDIS_ADDR=localhost:6379
+REDIS_PASSWORD=your_redis_password
+```
+
+### 网络配置
+
+如果 chat-deploy 的服务运行在其他 Docker 网络中，需要将可观测性组件加入该网络。
+
+编辑 `docker-compose-observability.yaml`，添加外部网络：
+
+```yaml
+networks:
+  chat-deploy-obs:
+    driver: bridge
+  # 添加外部网络
+  chat-deploy-network:
+    external: true
+
+services:
+  mongodb-exporter:
+    networks:
+      - chat-deploy-obs
+      - chat-deploy-network  # 加入 chat-deploy 网络
+```
+
+## 监控面板
+
+在 itom-platform 的 Grafana 中，使用 `chat-deploy-dashboard` 查看：
+
+- 服务健康状态
+- Go 运行时指标（CPU、内存、GC）
+- MongoDB 操作统计（query/insert/update/delete）
+- MongoDB 连接数
+- Redis 指标（如已配置）
+- 应用日志
+
+## 故障排查
+
+### 日志未采集
+
+1. 检查 Promtail 是否正常运行：
+   ```bash
+   docker logs chat-deploy-promtail
+   ```
+
+2. 确认 LOKI_URL 配置正确：
+   ```bash
+   echo $LOKI_URL
+   # 应输出类似：http://192.168.1.100/loki/api/v1/push
+   ```
+
+3. 测试网络连通性：
+   ```bash
+   curl -v http://<OBS_HOST>/loki/api/v1/push
+   ```
+
+### MongoDB 指标缺失
+
+1. 检查 MongoDB Exporter 是否正常运行：
+   ```bash
+   docker logs chat-deploy-mongodb-exporter
+   ```
+
+2. 验证 MongoDB 连接：
+   ```bash
+   curl http://localhost:9216/metrics | grep mongodb_up
+   # 应输出：mongodb_up 1
+   ```
+
+3. 确认 MongoDB URI 格式正确且网络可达
+
+### 指标未推送到中心
+
+1. 检查 Prometheus Agent 日志：
+   ```bash
+   docker logs chat-deploy-prometheus-agent
+   ```
+
+2. 确认 remote_write URL 配置正确
+
+3. 检查网络防火墙是否允许访问中心端口
+
+## 更新配置
+
+修改 `config.env` 后，重启服务：
+
+```bash
+docker compose --env-file config.env -f docker-compose-observability.yaml down
+docker compose --env-file config.env -f docker-compose-observability.yaml --profile metrics up -d
+```

deployments/observability/config.env.example

@@ -0,0 +1,71 @@
+# ==============================
+# chat-deploy 可观测性配置
+# ==============================
+
+# 项目标识（必须与 itom-platform 中配置一致）
+OBS_PROJECT=chat-deploy
+OBS_SERVICE=chat-deploy
+OBS_SERVICE_NAME=chat-deploy
+OBS_ENV=prod
+
+# ==============================
+# 中心侧连接配置（必须修改）
+# ==============================
+# itom-platform 可观测中心的地址
+OBS_HOST=CHANGE_ME
+OBS_SCHEME=http
+
+# 鉴权 token（必须与 itom-platform 中心侧一致）
+OBS_AUTH_TOKEN=CHANGE_ME
+
+# ==============================
+# 日志采集配置（Promtail）
+# ==============================
+# Loki 写入 URL（走网关）
+LOKI_URL=${OBS_SCHEME}://${OBS_HOST}/loki/api/v1/push
+
+# Promtail 镜像版本
+PROMTAIL_IMAGE=grafana/promtail:3.0.0
+
+# Docker API 版本（避免与旧 Docker daemon 协议不兼容）
+DOCKER_API_VERSION=1.44
+
+# ==============================
+# 指标采集配置（Prometheus Agent）
+# ==============================
+# 中心侧 remote_write receiver 地址
+METRICS_REMOTE_WRITE_URL=${OBS_SCHEME}://${OBS_HOST}/api/v1/write
+
+# 中心侧是否开启 remote_write 鉴权
+OBS_AUTH_ENABLE=false
+
+# 业务服务指标采集目标（可选，格式：name=host:port,name2=host:port）
+METRICS_TARGETS=
+
+# ==============================
+# MongoDB 配置（必须修改）
+# ==============================
+# MongoDB 连接 URI（用于 mongodb-exporter）
+# 格式：mongodb://username:password@host:port/database?authSource=admin
+MONGODB_URI=mongodb://openIM:openIM123@localhost:37017/openim_v3?authSource=openim_v3
+
+# MongoDB Exporter 采集目标（自动配置，通常无需修改）
+MONGODB_EXPORTER_TARGETS=mongodb-exporter:9216
+MONGODB_EXPORTER_SERVICE=chat-deploy
+
+# ==============================
+# Redis 配置（可选）
+# ==============================
+# Redis 地址（用于 redis-exporter）
+REDIS_ADDR=localhost:6379
+REDIS_PASSWORD=
+
+# Redis Exporter 采集目标
+REDIS_EXPORTER_TARGETS=redis-exporter:9121
+REDIS_EXPORTER_SERVICE=chat-deploy
+
+# ==============================
+# Node Exporter 配置（可选）
+# ==============================
+NODE_EXPORTER_TARGETS=node-exporter:9100
+NODE_EXPORTER_SERVICE=chat-deploy

deployments/observability/config/prometheus-agent-entrypoint.sh

@@ -0,0 +1,236 @@
+#!/usr/bin/env sh
+set -eu
+
+# ------------------------------
+# chat-deploy 指标采集：Prometheus Agent Entrypoint
+# - 根据环境变量生成 /prometheus/prometheus.yml 与 /prometheus/targets.json
+# - 然后以 agent 模式启动 Prometheus（remote_write 推送到 itom-platform 中心）
+#
+# 关键环境变量（来自 config.env）：
+# - METRICS_REMOTE_WRITE_URL=http(s)://<OBS_HOST>/api/v1/write
+# - METRICS_TARGETS=name=host:port,name2=host:port
+# - OBS_AUTH_ENABLE=false/true（中心侧是否要求鉴权）
+# - OBS_AUTH_TOKEN=xxxxx（当 OBS_AUTH_ENABLE=true 时必填）
+# - OBS_PROJECT/OBS_ENV：写入 labels，便于中心侧筛选
+# ------------------------------
+
+METRICS_REMOTE_WRITE_URL="${METRICS_REMOTE_WRITE_URL:-}"
+OBS_AUTH_ENABLE="${OBS_AUTH_ENABLE:-false}"
+OBS_AUTH_TOKEN="${OBS_AUTH_TOKEN:-}"
+METRICS_TARGETS="${METRICS_TARGETS:-}"
+
+OBS_PROJECT="${OBS_PROJECT:-chat-deploy}"
+OBS_ENV="${OBS_ENV:-prod}"
+OBS_SERVICE="${OBS_SERVICE:-chat-deploy}"
+OBS_SERVICE_NAME="${OBS_SERVICE_NAME:-$OBS_SERVICE}"
+if [ -z "$OBS_SERVICE_NAME" ]; then
+  OBS_SERVICE_NAME="chat-deploy"
+fi
+
+is_truthy() { case "$1" in 1|true|TRUE|yes|YES|on|ON) return 0 ;; esac; return 1; }
+
+if [ -z "$METRICS_REMOTE_WRITE_URL" ]; then
+  echo "[prometheus-agent] FAIL: METRICS_REMOTE_WRITE_URL 为空" >&2
+  exit 2
+fi
+
+if is_truthy "$OBS_AUTH_ENABLE" && [ -z "$OBS_AUTH_TOKEN" ]; then
+  echo "[prometheus-agent] FAIL: OBS_AUTH_ENABLE=true 但 OBS_AUTH_TOKEN 为空" >&2
+  exit 2
+fi
+
+# 确保数据目录存在且有写权限
+mkdir -p /prometheus/data
+chmod -R 777 /prometheus 2>/dev/null || true
+
+TARGETS_JSON="/prometheus/targets.json"
+echo "[" > "$TARGETS_JSON"
+
+first=1
+IFS=','
+for item in $METRICS_TARGETS; do
+  item="$(printf "%s" "$item" | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')"
+  [ -z "$item" ] && continue
+
+  name="$(printf "%s" "$item" | cut -d= -f1 | tr -d '[:space:]')"
+  target="$(printf "%s" "$item" | cut -d= -f2- | tr -d '[:space:]')"
+  [ -z "$name" ] && continue
+  [ -z "$target" ] && continue
+
+  target="$(printf "%s" "$target" | sed -E 's#^https?://##; s#/.*$##')"
+
+  if [ $first -eq 1 ]; then first=0; else echo "," >> "$TARGETS_JSON"; fi
+  cat >> "$TARGETS_JSON" <<EOF
+  {"targets":["$target"],"labels":{"service":"$name","project":"$OBS_PROJECT","env":"$OBS_ENV"}}
+EOF
+done
+unset IFS
+echo "]" >> "$TARGETS_JSON"
+
+if [ "$first" -eq 1 ]; then
+  echo "[prometheus-agent] WARN: METRICS_TARGETS 为空或格式无效，仅采集 prometheus-agent 自身指标和 Exporter 指标" >&2
+fi
+
+CONFIG="/prometheus/prometheus.yml"
+cat > "$CONFIG" <<EOF
+global:
+  scrape_interval: 15s
+  external_labels:
+    project: $OBS_PROJECT
+    env: $OBS_ENV
+    service: $OBS_SERVICE_NAME
+
+scrape_configs:
+  # Prometheus Agent 自身指标（用于验证采集链路是否正常）
+  - job_name: "prometheus-agent"
+    static_configs:
+      - targets: ["localhost:9090"]
+
+  # 业务服务指标（通过 METRICS_TARGETS 或 file_sd_configs 配置）
+  - job_name: "services"
+    metrics_path: /metrics
+    file_sd_configs:
+      - files: ["$TARGETS_JSON"]
+        refresh_interval: 10s
+EOF
+
+# 自动发现并采集 Exporter 指标（Redis、MongoDB）
+echo "" >> "$CONFIG"
+echo "  # 自动发现/手动指定的 Exporter 指标" >> "$CONFIG"
+
+trim() {
+  printf "%s" "$1" | sed 's/^[[:space:]]*//;s/[[:space:]]*$//'
+}
+
+normalize_target() {
+  raw="$1"
+  default_port="$2"
+  raw="$(printf "%s" "$raw" | sed -E 's#^https?://##; s#/.*$##')"
+  raw="$(trim "$raw")"
+  [ -z "$raw" ] && return 1
+  case "$raw" in
+    *:*) printf "%s" "$raw" ;;
+    *) printf "%s:%s" "$raw" "$default_port" ;;
+  esac
+}
+
+append_exporter_job() {
+  job="$1"
+  targets="$2"
+  default_port="$3"
+  service_label="$4"
+  targets_list=""
+
+  IFS=','
+  for item in $targets; do
+    item="$(normalize_target "$item" "$default_port" || true)"
+    [ -z "$item" ] && continue
+    if [ -z "$targets_list" ]; then
+      targets_list="$item"
+    else
+      targets_list="${targets_list},$item"
+    fi
+  done
+  unset IFS
+
+  if [ -z "$targets_list" ]; then
+    return 0
+  fi
+
+  echo "  - job_name: '$job'" >> "$CONFIG"
+  echo "    static_configs:" >> "$CONFIG"
+  echo "      - targets:" >> "$CONFIG"
+  IFS=','
+  for item in $targets_list; do
+    echo "          - '$item'" >> "$CONFIG"
+  done
+  unset IFS
+  cat >> "$CONFIG" <<EOF
+        labels:
+          project: '${OBS_PROJECT}'
+          service: '${service_label}'
+EOF
+}
+
+# Redis Exporter（端口 9121）
+REDIS_EXPORTER_HOST=""
+REDIS_EXPORTER_TARGETS="${REDIS_EXPORTER_TARGETS:-}"
+REDIS_EXPORTER_SERVICE="${REDIS_EXPORTER_SERVICE:-$OBS_SERVICE_NAME}"
+if getent hosts redis-exporter >/dev/null 2>&1; then
+  REDIS_EXPORTER_HOST="redis-exporter"
+elif getent hosts chat-deploy-redis-exporter >/dev/null 2>&1; then
+  REDIS_EXPORTER_HOST="chat-deploy-redis-exporter"
+fi
+if [ -n "$REDIS_EXPORTER_HOST" ]; then
+  if [ -n "$REDIS_EXPORTER_TARGETS" ]; then
+    REDIS_EXPORTER_TARGETS="${REDIS_EXPORTER_TARGETS},${REDIS_EXPORTER_HOST}:9121"
+  else
+    REDIS_EXPORTER_TARGETS="${REDIS_EXPORTER_HOST}:9121"
+  fi
+fi
+append_exporter_job "redis" "$REDIS_EXPORTER_TARGETS" "9121" "$REDIS_EXPORTER_SERVICE"
+if [ -n "$REDIS_EXPORTER_TARGETS" ]; then
+  echo "[prometheus-agent] Redis Exporter 采集目标已配置（project=${OBS_PROJECT} service=${REDIS_EXPORTER_SERVICE}）"
+fi
+
+# MongoDB Exporter（端口 9216）- chat-deploy 使用 MongoDB
+MONGODB_EXPORTER_HOST=""
+MONGODB_EXPORTER_TARGETS="${MONGODB_EXPORTER_TARGETS:-}"
+MONGODB_EXPORTER_SERVICE="${MONGODB_EXPORTER_SERVICE:-$OBS_SERVICE_NAME}"
+if getent hosts mongodb-exporter >/dev/null 2>&1; then
+  MONGODB_EXPORTER_HOST="mongodb-exporter"
+elif getent hosts chat-deploy-mongodb-exporter >/dev/null 2>&1; then
+  MONGODB_EXPORTER_HOST="chat-deploy-mongodb-exporter"
+fi
+if [ -n "$MONGODB_EXPORTER_HOST" ]; then
+  if [ -n "$MONGODB_EXPORTER_TARGETS" ]; then
+    MONGODB_EXPORTER_TARGETS="${MONGODB_EXPORTER_TARGETS},${MONGODB_EXPORTER_HOST}:9216"
+  else
+    MONGODB_EXPORTER_TARGETS="${MONGODB_EXPORTER_HOST}:9216"
+  fi
+fi
+append_exporter_job "mongodb" "$MONGODB_EXPORTER_TARGETS" "9216" "$MONGODB_EXPORTER_SERVICE"
+if [ -n "$MONGODB_EXPORTER_TARGETS" ]; then
+  echo "[prometheus-agent] MongoDB Exporter 采集目标已配置（project=${OBS_PROJECT} service=${MONGODB_EXPORTER_SERVICE}）"
+fi
+
+# Node Exporter（端口 9100）- 用于系统级指标（CPU/Memory/Disk/Network/IO）
+NODE_EXPORTER_HOST=""
+NODE_EXPORTER_TARGETS="${NODE_EXPORTER_TARGETS:-}"
+NODE_EXPORTER_SERVICE="${NODE_EXPORTER_SERVICE:-$OBS_SERVICE_NAME}"
+if getent hosts node-exporter >/dev/null 2>&1; then
+  NODE_EXPORTER_HOST="node-exporter"
+elif getent hosts chat-deploy-node-exporter >/dev/null 2>&1; then
+  NODE_EXPORTER_HOST="chat-deploy-node-exporter"
+fi
+if [ -n "$NODE_EXPORTER_HOST" ]; then
+  if [ -n "$NODE_EXPORTER_TARGETS" ]; then
+    NODE_EXPORTER_TARGETS="${NODE_EXPORTER_TARGETS},${NODE_EXPORTER_HOST}:9100"
+  else
+    NODE_EXPORTER_TARGETS="${NODE_EXPORTER_HOST}:9100"
+  fi
+fi
+append_exporter_job "node" "$NODE_EXPORTER_TARGETS" "9100" "$NODE_EXPORTER_SERVICE"
+if [ -n "$NODE_EXPORTER_TARGETS" ]; then
+  echo "[prometheus-agent] Node Exporter 采集目标已配置（project=${OBS_PROJECT} service=${NODE_EXPORTER_SERVICE}）"
+fi
+
+cat >> "$CONFIG" <<EOF
+
+remote_write:
+  - url: "$METRICS_REMOTE_WRITE_URL"
+EOF
+
+if is_truthy "$OBS_AUTH_ENABLE"; then
+  echo "    bearer_token: \"$OBS_AUTH_TOKEN\"" >> "$CONFIG"
+  echo "[prometheus-agent] remote_write 鉴权已启用" >&2
+else
+  echo "[prometheus-agent] remote_write 鉴权未启用" >&2
+fi
+
+echo "[prometheus-agent] 配置文件已生成:"
+cat "$CONFIG"
+echo ""
+
+# Prometheus 3.x 不再需要 --enable-feature=agent
+exec /bin/prometheus --config.file=/prometheus/prometheus.yml --storage.tsdb.path=/prometheus/data --web.enable-lifecycle

deployments/observability/config/promtail.yaml

@@ -0,0 +1,61 @@
+server:
+  http_listen_port: 9080
+  grpc_listen_port: 0
+
+positions:
+  filename: /tmp/positions.yaml
+
+clients:
+  - url: ${LOKI_URL}
+    bearer_token: ${OBS_AUTH_TOKEN}
+
+scrape_configs:
+  # ============================================
+  # chat-deploy 业务层日志采集
+  # ============================================
+  - job_name: chat-deploy-logs
+    static_configs:
+      - targets:
+          - localhost
+        labels:
+          job: chat-deploy-logs
+          project: ${OBS_PROJECT}
+          service: ${OBS_SERVICE}
+          log_layer: business
+          __path__: /var/lib/docker/containers/*/*-json.log
+
+    pipeline_stages:
+      # 解析 Docker JSON 日志格式
+      - docker: {}
+
+      # 从文件路径提取容器 ID
+      - regex:
+          source: filename
+          expression: '/var/lib/docker/containers/(?P<container_id>[0-9a-f]{12})[0-9a-f]*/'
+      - labels:
+          container_id:
+
+      # ============================================
+      # 过滤规则：只丢弃基础设施层日志
+      # 保留 chat-deploy 业务层日志（包括Go服务的caller日志）
+      # ============================================
+      
+      # 1. 丢弃 promtail/prometheus/grafana 等基础设施组件的内部日志
+      - drop:
+          expression: 'caller=.*(promtail|prometheus|grafana|loki|exporter).*\.go:[0-9]+'
+          drop_counter_reason: infra_go_internal_log
+
+      # 2. 丢弃基础设施组件日志（prometheus/grafana/exporter等）
+      - drop:
+          expression: '(component=|target=|scrape_pool=|instance=.*exporter)'
+          drop_counter_reason: infrastructure_component
+
+      # 3. 丢弃 Docker 服务发现和 API 错误日志
+      - drop:
+          expression: '(docker_discovery|Unable to refresh target groups|client version.*is too old|Minimum supported API version)'
+          drop_counter_reason: docker_api_error
+
+      # 4. 丢弃文件监控事件日志（promtail 内部）
+      - drop:
+          expression: '(file watcher event|filetargetmanager|fsnotify)'
+          drop_counter_reason: file_watcher

deployments/observability/docker-compose-observability.yaml

@@ -0,0 +1,135 @@
+# chat-deploy 可观测性组件部署配置
+# 用于向 itom-platform 可观测中心推送日志和指标
+#
+# 使用方法：
+#   1. 复制 config.env.example 为 config.env 并修改配置
+#   2. 启动服务：docker compose --env-file config.env -f docker-compose-observability.yaml up -d
+#   3. 如需启用指标采集：docker compose --env-file config.env -f docker-compose-observability.yaml --profile metrics up -d
+
+services:
+  # ==============================
+  # 日志采集（Promtail）
+  # 采集 Docker 容器日志并推送到 itom-platform 的 Loki
+  # ==============================
+  promtail:
+    image: "${PROMTAIL_IMAGE:-grafana/promtail:3.0.0}"
+    container_name: chat-deploy-promtail
+    restart: always
+    user: "0"
+    command: ["-config.file=/etc/promtail/config.yml", "-config.expand-env=true"]
+    # 禁用 promtail 自身的日志输出到 Docker，避免日志循环
+    logging:
+      driver: "none"
+    environment:
+      - LOKI_URL=${LOKI_URL}
+      - OBS_AUTH_TOKEN=${OBS_AUTH_TOKEN}
+      - OBS_PROJECT=${OBS_PROJECT}
+      - OBS_SERVICE=${OBS_SERVICE}
+      - DOCKER_API_VERSION=${DOCKER_API_VERSION:-1.44}
+    volumes:
+      - /var/lib/docker/containers:/var/lib/docker/containers:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+      - ./config/promtail.yaml:/etc/promtail/config.yml:ro
+    networks:
+      - chat-deploy-obs
+
+  # ==============================
+  # 指标采集（Prometheus Agent Remote Write）
+  # 可选，需要 --profile metrics 启用
+  # ==============================
+  prometheus-agent:
+    profiles: ["metrics"]
+    image: prom/prometheus:latest
+    container_name: chat-deploy-prometheus-agent
+    restart: always
+    user: "0"
+    command:
+      - "--config.file=/prometheus/prometheus.yml"
+      - "--storage.tsdb.path=/prometheus"
+    environment:
+      - METRICS_REMOTE_WRITE_URL=${METRICS_REMOTE_WRITE_URL}
+      - METRICS_TARGETS=${METRICS_TARGETS}
+      - OBS_AUTH_ENABLE=${OBS_AUTH_ENABLE:-false}
+      - OBS_AUTH_TOKEN=${OBS_AUTH_TOKEN}
+      - OBS_PROJECT=${OBS_PROJECT}
+      - OBS_SERVICE=${OBS_SERVICE}
+      - OBS_SERVICE_NAME=${OBS_SERVICE_NAME}
+      - OBS_ENV=${OBS_ENV:-prod}
+      - REDIS_EXPORTER_TARGETS=${REDIS_EXPORTER_TARGETS}
+      - MONGODB_EXPORTER_TARGETS=${MONGODB_EXPORTER_TARGETS}
+      - NODE_EXPORTER_TARGETS=${NODE_EXPORTER_TARGETS}
+      - REDIS_EXPORTER_SERVICE=${REDIS_EXPORTER_SERVICE}
+      - MONGODB_EXPORTER_SERVICE=${MONGODB_EXPORTER_SERVICE}
+      - NODE_EXPORTER_SERVICE=${NODE_EXPORTER_SERVICE}
+    volumes:
+      - prometheus_agent_data:/prometheus
+      - ./config/prometheus-agent-entrypoint.sh:/etc/prometheus/entrypoint.sh:ro
+    entrypoint: ["/bin/sh", "/etc/prometheus/entrypoint.sh"]
+    networks:
+      - chat-deploy-obs
+    depends_on:
+      - mongodb-exporter
+
+  # ==============================
+  # MongoDB Exporter
+  # 采集 MongoDB 指标，需要 --profile metrics 启用
+  # ==============================
+  mongodb-exporter:
+    profiles: ["metrics"]
+    image: percona/mongodb_exporter:0.40.0
+    container_name: chat-deploy-mongodb-exporter
+    restart: always
+    command:
+      - "--mongodb.uri=${MONGODB_URI}"
+      - "--compatible-mode"
+      - "--collect-all"
+    environment:
+      - MONGODB_URI=${MONGODB_URI}
+    ports:
+      - "9216:9216"
+    networks:
+      - chat-deploy-obs
+
+  # ==============================
+  # Redis Exporter（可选）
+  # 采集 Redis 指标，需要 --profile metrics 启用
+  # ==============================
+  redis-exporter:
+    profiles: ["metrics"]
+    image: oliver006/redis_exporter:latest
+    container_name: chat-deploy-redis-exporter
+    restart: always
+    environment:
+      - REDIS_ADDR=${REDIS_ADDR}
+      - REDIS_PASSWORD=${REDIS_PASSWORD}
+    ports:
+      - "9121:9121"
+    networks:
+      - chat-deploy-obs
+
+  # ==============================
+  # Node Exporter（可选）
+  # 采集系统级指标（CPU/Memory/Disk/Network）
+  # 需要 --profile metrics 启用
+  # ==============================
+  node-exporter:
+    profiles: ["metrics"]
+    image: prom/node-exporter:latest
+    container_name: chat-deploy-node-exporter
+    restart: always
+    command:
+      - "--path.rootfs=/host"
+      - "--collector.filesystem.mount-points-exclude=^/(sys|proc|dev|host|etc)($$|/)"
+    volumes:
+      - /:/host:ro,rslave
+    ports:
+      - "9100:9100"
+    networks:
+      - chat-deploy-obs
+
+networks:
+  chat-deploy-obs:
+    driver: bridge
+
+volumes:
+  prometheus_agent_data: {}