[#33] 使用新构建agent二进制发布包

refs #29

.gitattributes

@@ -0,0 +1 @@
+src/metric/client-plugins/all-in-one-full/plugins/*/bin/* filter=lfs diff=lfs merge=lfs -text

build/build_images.sh

@@ -10,19 +10,29 @@ Usage: $0 [OPTIONS]
 Options:
   --intranet        Use intranet mirror for log/bind builds
   --master-offline  Build master offline image (requires src/master/offline_wheels.tar.gz)
+  --metric          Build metric module images (ftp, prometheus, grafana, test nodes)
+  --no-cache        Build all images without using Docker layer cache
+  --only LIST       Comma-separated targets to build: core,master,metric,web,alert,sys,all
   -h, --help        Show this help message
 
 Examples:
   $0                             # Build with default sources
   $0 --intranet                  # Build with intranet mirror
   $0 --master-offline            # Additionally build argus-master:offline
-  $0 --intranet --master-offline
+  $0 --metric                    # Additionally build metric module images
+  $0 --intranet --master-offline --metric
 EOF
 }
 
 use_intranet=false
+build_core=true
 build_master=true
 build_master_offline=false
+build_metric=true
+build_web=true
+build_alert=true
+build_sys=true
+no_cache=false
 
 while [[ $# -gt 0 ]]; do
   case $1 in
@@ -39,6 +49,35 @@ while [[ $# -gt 0 ]]; do
       build_master_offline=true
       shift
       ;;
+    --metric)
+      build_metric=true
+      shift
+      ;;
+    --no-cache)
+      no_cache=true
+      shift
+      ;;
+    --only)
+      if [[ -z ${2:-} ]]; then
+        echo "--only requires a target list" >&2; exit 1
+      fi
+      sel="$2"; shift 2
+      # reset all, then enable selected
+      build_core=false; build_master=false; build_metric=false; build_web=false; build_alert=false; build_sys=false
+      IFS=',' read -ra parts <<< "$sel"
+      for p in "${parts[@]}"; do
+        case "$p" in
+          core) build_core=true ;;
+          master) build_master=true ;;
+          metric) build_metric=true ;;
+          web) build_web=true ;;
+          alert) build_alert=true ;;
+          sys) build_sys=true ;;
+          all) build_core=true; build_master=true; build_metric=true; build_web=true; build_alert=true; build_sys=true ;;
+          *) echo "Unknown --only target: $p" >&2; exit 1 ;;
+        esac
+      done
+      ;;
     -h|--help)
       show_help
       exit 0
@@ -65,6 +104,10 @@ cd "$root"
 load_build_user
 build_args+=("--build-arg" "ARGUS_BUILD_UID=${ARGUS_BUILD_UID}" "--build-arg" "ARGUS_BUILD_GID=${ARGUS_BUILD_GID}")
 
+if [[ "$no_cache" == true ]]; then
+  build_args+=("--no-cache")
+fi
+
 master_root="$root/src/master"
 master_offline_tar="$master_root/offline_wheels.tar.gz"
 master_offline_dir="$master_root/offline_wheels"
@@ -105,14 +148,22 @@ build_image() {
   local image_name=$1
   local dockerfile_path=$2
   local tag=$3
+  local context="."
   shift 3
+
+  if [[ $# -gt 0 ]]; then
+    context=$1
+    shift
+  fi
+
   local extra_args=("$@")
 
   echo "🔄 Building $image_name image..."
   echo "   Dockerfile: $dockerfile_path"
   echo "   Tag: $tag"
+  echo "   Context: $context"
 
-  if docker build "${build_args[@]}" "${extra_args[@]}" -f "$dockerfile_path" -t "$tag" .; then
+  if docker build "${build_args[@]}" "${extra_args[@]}" -f "$dockerfile_path" -t "$tag" "$context"; then
     echo "✅ $image_name image built successfully"
     return 0
   else
@@ -121,29 +172,59 @@ build_image() {
   fi
 }
 
+pull_base_image() {
+  local image_ref=$1
+  local attempts=${2:-3}
+  local delay=${3:-5}
+
+  # If the image already exists locally, skip pulling.
+  if docker image inspect "$image_ref" >/dev/null 2>&1; then
+    echo "   Local image present; skip pull: $image_ref"
+    return 0
+  fi
+
+  for ((i=1; i<=attempts; i++)); do
+    echo "   Pulling base image ($i/$attempts): $image_ref"
+    if docker pull "$image_ref" >/dev/null; then
+      echo "   Base image ready: $image_ref"
+      return 0
+    fi
+    echo "   Pull failed: $image_ref"
+    if (( i < attempts )); then
+      echo "   Retrying in ${delay}s..."
+      sleep "$delay"
+    fi
+  done
+
+  echo "❌ Unable to pull base image after ${attempts} attempts: $image_ref"
+  return 1
+}
+
 images_built=()
 build_failed=false
 
-if build_image "Elasticsearch" "src/log/elasticsearch/build/Dockerfile" "argus-elasticsearch:latest"; then
-  images_built+=("argus-elasticsearch:latest")
-else
-  build_failed=true
-fi
+if [[ "$build_core" == true ]]; then
+  if build_image "Elasticsearch" "src/log/elasticsearch/build/Dockerfile" "argus-elasticsearch:latest"; then
+    images_built+=("argus-elasticsearch:latest")
+  else
+    build_failed=true
+  fi
 
-echo ""
+  echo ""
 
-if build_image "Kibana" "src/log/kibana/build/Dockerfile" "argus-kibana:latest"; then
-  images_built+=("argus-kibana:latest")
-else
-  build_failed=true
-fi
+  if build_image "Kibana" "src/log/kibana/build/Dockerfile" "argus-kibana:latest"; then
+    images_built+=("argus-kibana:latest")
+  else
+    build_failed=true
+  fi
 
-echo ""
+  echo ""
 
-if build_image "BIND9" "src/bind/build/Dockerfile" "argus-bind9:latest"; then
-  images_built+=("argus-bind9:latest")
-else
-  build_failed=true
+  if build_image "BIND9" "src/bind/build/Dockerfile" "argus-bind9:latest"; then
+    images_built+=("argus-bind9:latest")
+  else
+    build_failed=true
+  fi
 fi
 
 echo ""
@@ -159,6 +240,9 @@ if [[ "$build_master" == true ]]; then
   if [[ "$build_master_offline" == true ]]; then
     master_args+=("--offline")
   fi
+  if [[ "$no_cache" == true ]]; then
+    master_args+=("--no-cache")
+  fi
   if ./scripts/build_images.sh "${master_args[@]}"; then
     if [[ "$build_master_offline" == true ]]; then
       images_built+=("argus-master:offline")
@@ -171,6 +255,127 @@ if [[ "$build_master" == true ]]; then
   popd >/dev/null
 fi
 
+if [[ "$build_metric" == true ]]; then
+  echo ""
+  echo "Building Metric module images..."
+
+  metric_base_images=(
+    "ubuntu:22.04"
+    "ubuntu/prometheus:3-24.04_stable"
+    "grafana/grafana:11.1.0"
+  )
+
+  for base_image in "${metric_base_images[@]}"; do
+    if ! pull_base_image "$base_image"; then
+      build_failed=true
+    fi
+  done
+
+  metric_builds=(
+    "Metric FTP|src/metric/ftp/build/Dockerfile|argus-metric-ftp:latest|src/metric/ftp/build"
+    "Metric Prometheus|src/metric/prometheus/build/Dockerfile|argus-metric-prometheus:latest|src/metric/prometheus/build"
+    "Metric Grafana|src/metric/grafana/build/Dockerfile|argus-metric-grafana:latest|src/metric/grafana/build"
+  )
+
+  for build_spec in "${metric_builds[@]}"; do
+    IFS='|' read -r image_label dockerfile_path image_tag build_context <<< "$build_spec"
+    if build_image "$image_label" "$dockerfile_path" "$image_tag" "$build_context"; then
+      images_built+=("$image_tag")
+    else
+      build_failed=true
+    fi
+    echo ""
+  done
+fi
+
+# =======================================
+# Sys (system tests) node images
+# =======================================
+
+if [[ "$build_sys" == true ]]; then
+  echo ""
+  echo "Building Sys node images..."
+
+  sys_base_images=(
+    "ubuntu:22.04"
+    "nvidia/cuda:12.2.2-runtime-ubuntu22.04"
+  )
+
+  for base_image in "${sys_base_images[@]}"; do
+    if ! pull_base_image "$base_image"; then
+      build_failed=true
+    fi
+  done
+
+  sys_builds=(
+    "Sys Node|src/sys/build/node/Dockerfile|argus-sys-node:latest|."
+    "Sys Metric Test Node|src/sys/build/test-node/Dockerfile|argus-sys-metric-test-node:latest|."
+    "Sys Metric Test GPU Node|src/sys/build/test-gpu-node/Dockerfile|argus-sys-metric-test-gpu-node:latest|."
+  )
+
+  for build_spec in "${sys_builds[@]}"; do
+    IFS='|' read -r image_label dockerfile_path image_tag build_context <<< "$build_spec"
+    if build_image "$image_label" "$dockerfile_path" "$image_tag" "$build_context"; then
+      images_built+=("$image_tag")
+    else
+      build_failed=true
+    fi
+    echo ""
+  done
+fi
+
+# =======================================
+# Web & Alert module images
+# =======================================
+
+if [[ "$build_web" == true || "$build_alert" == true ]]; then
+  echo ""
+  echo "Building Web and Alert module images..."
+
+  # Pre-pull commonly used base images for stability
+  web_alert_base_images=(
+    "node:20"
+    "ubuntu:24.04"
+  )
+
+  for base_image in "${web_alert_base_images[@]}"; do
+    if ! pull_base_image "$base_image"; then
+      build_failed=true
+    fi
+  done
+
+  if [[ "$build_web" == true ]]; then
+    web_builds=(
+      "Web Frontend|src/web/build_tools/frontend/Dockerfile|argus-web-frontend:latest|."
+      "Web Proxy|src/web/build_tools/proxy/Dockerfile|argus-web-proxy:latest|."
+    )
+    for build_spec in "${web_builds[@]}"; do
+      IFS='|' read -r image_label dockerfile_path image_tag build_context <<< "$build_spec"
+      if build_image "$image_label" "$dockerfile_path" "$image_tag" "$build_context"; then
+        images_built+=("$image_tag")
+      else
+        build_failed=true
+      fi
+      echo ""
+    done
+  fi
+
+  if [[ "$build_alert" == true ]]; then
+    alert_builds=(
+      "Alertmanager|src/alert/alertmanager/build/Dockerfile|argus-alertmanager:latest|."
+    )
+    for build_spec in "${alert_builds[@]}"; do
+      IFS='|' read -r image_label dockerfile_path image_tag build_context <<< "$build_spec"
+      if build_image "$image_label" "$dockerfile_path" "$image_tag" "$build_context"; then
+        images_built+=("$image_tag")
+      else
+        build_failed=true
+      fi
+      echo ""
+    done
+  fi
+fi
+
 echo "======================================="
 echo "📦 Build Summary"
 echo "======================================="
@@ -197,7 +402,6 @@ if [[ "$build_master_offline" == true ]]; then
   echo ""
   echo "🧳 Master offline wheels 已解压到 $master_offline_dir"
 fi
-
 echo ""
 echo "🚀 Next steps:"
 echo "   ./build/save_images.sh --compress          # 导出镜像"

build/save_images.sh

@@ -68,6 +68,12 @@ declare -A images=(
     ["argus-kibana:latest"]="argus-kibana-latest.tar"
     ["argus-bind9:latest"]="argus-bind9-latest.tar"
     ["argus-master:offline"]="argus-master-offline.tar"
+    ["argus-metric-ftp:latest"]="argus-metric-ftp-latest.tar"
+    ["argus-metric-prometheus:latest"]="argus-metric-prometheus-latest.tar"
+    ["argus-metric-grafana:latest"]="argus-metric-grafana-latest.tar"
+    ["argus-web-frontend:latest"]="argus-web-frontend-latest.tar"
+    ["argus-web-proxy:latest"]="argus-web-proxy-latest.tar"
+    ["argus-alertmanager:latest"]="argus-alertmanager-latest.tar"
 )
 
 # 函数：检查镜像是否存在

src/agent/README.md

@@ -34,6 +34,18 @@ Agent 不再依赖配置文件；所有参数均由环境变量与主机名推
 | `MASTER_ENDPOINT` | 是 | N/A | Master 基础地址，可写 `http://host:3000` 或 `host:3000`（自动补全 `http://`）。 |
 | `REPORT_INTERVAL_SECONDS` | 否 | `60` | 状态上报间隔（秒）。必须为正整数。 |
 | `AGENT_HOSTNAME` | 否 | `$(hostname)` | 覆盖容器内主机名，便于测试或特殊命名需求。 |
+| `AGENT_ENV` | 否 | 来源于主机名 | 运行环境标识（如 `dev`、`prod`）。与 `AGENT_USER`、`AGENT_INSTANCE` 必须同时设置。 |
+| `AGENT_USER` | 否 | 来源于主机名 | 归属用户或团队标识。与 `AGENT_ENV`、`AGENT_INSTANCE` 必须同时设置。 |
+| `AGENT_INSTANCE` | 否 | 来源于主机名 | 实例编号或别名。与 `AGENT_ENV`、`AGENT_USER` 必须同时设置。 |
+
+主机名与元数据的解析优先级：
+
+1. 若设置 `AGENT_ENV` / `AGENT_USER` / `AGENT_INSTANCE` 且全部存在，则直接使用这些值。
+2. 否则检查历史 `node.json`（注册成功后由 Master 返回的信息），若包含 `env` / `user` / `instance` 则沿用。
+3. 若以上均不可用，则按历史约定从主机名解析 `env-user-instance` 前缀。
+4. 如果仍无法得到完整结果，Agent 启动会失败并提示需要提供上述环境变量。
+
+> 提示：在首次部署时需确保环境变量或主机名能够提供完整信息。完成注册后，Agent 会把 Master 返回的元数据写入 `node.json`，后续重启无需再次提供环境变量就能保持一致性。
 
 派生路径：
 

src/agent/app/collector.py

@@ -18,13 +18,12 @@ _HOSTNAME_PATTERN = re.compile(r"^([^-]+)-([^-]+)-([^-]+)-.*$")
 def collect_metadata(config: AgentConfig) -> Dict[str, Any]:
     """汇总节点注册需要的静态信息。"""
     hostname = config.hostname
-    env, user, instance = _parse_hostname(hostname)
     meta = {
         "hostname": hostname,
         "ip": _detect_ip_address(),
-        "env": env,
-        "user": user,
-        "instance": instance,
+        "env": config.environment,
+        "user": config.user,
+        "instance": config.instance,
         "cpu_number": _detect_cpu_count(),
         "memory_in_bytes": _detect_memory_bytes(),
         "gpu_number": _detect_gpu_count(),

src/agent/app/config.py

@@ -6,14 +6,21 @@ from dataclasses import dataclass
 from pathlib import Path
 from typing import Final
 
+from .state import load_node_state
 from .version import VERSION
+from .log import get_logger
 
 DEFAULT_REPORT_INTERVAL_SECONDS: Final[int] = 60
 
+LOGGER = get_logger("argus.agent.config")
+
 
 @dataclass(frozen=True)
 class AgentConfig:
     hostname: str
+    environment: str
+    user: str
+    instance: str
     node_file: str
     version: str
     master_endpoint: str
@@ -47,11 +54,68 @@ def _resolve_hostname() -> str:
     return os.environ.get("AGENT_HOSTNAME") or socket.gethostname()
 
 
+def _load_metadata_from_state(node_file: str) -> tuple[str, str, str] | None:
+    state = load_node_state(node_file)
+    if not state:
+        return None
+
+    meta = state.get("meta_data") or {}
+    env = meta.get("env") or state.get("env")
+    user = meta.get("user") or state.get("user")
+    instance = meta.get("instance") or state.get("instance")
+
+    if env and user and instance:
+        LOGGER.debug("Metadata resolved from node state", extra={"node_file": node_file})
+        return env, user, instance
+
+    LOGGER.warning(
+        "node.json missing metadata fields; ignoring",
+        extra={"node_file": node_file, "meta_data": meta},
+    )
+    return None
+
+
+def _resolve_metadata_fields(hostname: str, node_file: str) -> tuple[str, str, str]:
+    env = os.environ.get("AGENT_ENV")
+    user = os.environ.get("AGENT_USER")
+    instance = os.environ.get("AGENT_INSTANCE")
+
+    if env and user and instance:
+        return env, user, instance
+
+    if any([env, user, instance]):
+        LOGGER.warning(
+            "Incomplete metadata environment variables; falling back to persisted metadata",
+            extra={
+                "has_env": bool(env),
+                "has_user": bool(user),
+                "has_instance": bool(instance),
+            },
+        )
+
+    state_metadata = _load_metadata_from_state(node_file)
+    if state_metadata is not None:
+        return state_metadata
+
+    from .collector import _parse_hostname  # Local import to avoid circular dependency
+
+    env, user, instance = _parse_hostname(hostname)
+
+    if not all([env, user, instance]):
+        raise ValueError(
+            "Failed to determine metadata fields; set AGENT_ENV/USER/INSTANCE or use supported hostname pattern"
+        )
+
+    return env, user, instance
+
+
 def load_config() -> AgentConfig:
     """从环境变量推导配置，移除了外部配置文件依赖。"""
 
     hostname = _resolve_hostname()
     node_file = f"/private/argus/agent/{hostname}/node.json"
+    environment, user, instance = _resolve_metadata_fields(hostname, node_file)
+
     health_dir = f"/private/argus/agent/{hostname}/health/"
 
     master_endpoint_env = os.environ.get("MASTER_ENDPOINT")
@@ -66,6 +130,9 @@ def load_config() -> AgentConfig:
 
     return AgentConfig(
         hostname=hostname,
+        environment=environment,
+        user=user,
+        instance=instance,
         node_file=node_file,
         version=VERSION,
         master_endpoint=master_endpoint,

src/agent/scripts/build_binary.sh

@@ -12,6 +12,8 @@ VENV_DIR="$BUILD_ROOT/venv"
 
 AGENT_BUILD_IMAGE="${AGENT_BUILD_IMAGE:-python:3.11-slim-bullseye}"
 AGENT_BUILD_USE_DOCKER="${AGENT_BUILD_USE_DOCKER:-1}"
+# 默认在容器内忽略代理以避免公司内网代理在 Docker 网络不可达导致 pip 失败（可用 0 关闭）
+AGENT_BUILD_IGNORE_PROXY="${AGENT_BUILD_IGNORE_PROXY:-1}"
 USED_DOCKER=0
 
 run_host_build() {
@@ -71,6 +73,7 @@ run_docker_build() {
   pass_env_if_set http_proxy
   pass_env_if_set https_proxy
   pass_env_if_set no_proxy
+  pass_env_if_set AGENT_BUILD_IGNORE_PROXY
 
 build_script=$(cat <<'INNER'
 set -euo pipefail
@@ -82,6 +85,10 @@ rm -rf build dist
 mkdir -p build/pyinstaller dist
 python3 -m venv --copies build/venv
 source build/venv/bin/activate
+# 若指定忽略代理，则清空常见代理与 pip 镜像环境变量，避免容器内代理不可达
+if [ "${AGENT_BUILD_IGNORE_PROXY:-1}" = "1" ]; then
+  unset http_proxy https_proxy HTTP_PROXY HTTPS_PROXY PIP_INDEX_URL PIP_EXTRA_INDEX_URL PIP_TRUSTED_HOST
+fi
 pip install --upgrade pip
 pip install .
 pip install pyinstaller==6.6.0

src/agent/tests/docker-compose.yml

@@ -60,6 +60,36 @@ services:
         ipv4_address: 172.28.0.20
     restart: always
 
+  agent_env:
+    image: ubuntu:22.04
+    container_name: argus-agent-env-e2e
+    hostname: host_abc
+    depends_on:
+      - master
+      - bind
+    environment:
+      - MASTER_ENDPOINT=http://master.argus.com:3000
+      - REPORT_INTERVAL_SECONDS=2
+      - AGENT_ENV=prod
+      - AGENT_USER=ml
+      - AGENT_INSTANCE=node-3
+      - AGENT_HOSTNAME=host_abc
+      - "ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}"
+      - "ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}"
+    volumes:
+      - ./private/argus/agent/host_abc:/private/argus/agent/host_abc
+      - ./private/argus/agent/host_abc/health:/private/argus/agent/host_abc/health
+      - ./private/argus/etc:/private/argus/etc
+      - ../dist/argus-agent:/usr/local/bin/argus-agent:ro
+      - ./scripts/agent_entrypoint.sh:/usr/local/bin/agent-entrypoint.sh:ro
+      - ../scripts/agent_deployment_verify.sh:/usr/local/bin/agent_deployment_verify.sh:ro
+    entrypoint:
+      - /usr/local/bin/agent-entrypoint.sh
+    networks:
+      default:
+        ipv4_address: 172.28.0.21
+    restart: always
+
 networks:
   default:
     driver: bridge

src/agent/tests/scripts/00_e2e_test.sh

@@ -7,10 +7,10 @@ SCRIPTS=(
   "02_up.sh"
   "03_wait_and_assert_registration.sh"
   "04_write_health_files.sh"
-  "08_verify_agent.sh"
-  "05_assert_status_on_master.sh"
-  "06_restart_agent_and_reregister.sh"
-  "07_down.sh"
+  "05_verify_agent.sh"
+  "06_assert_status_on_master.sh"
+  "07_restart_agent_and_reregister.sh"
+  "08_down.sh"
 )
 
 for script in "${SCRIPTS[@]}"; do

src/agent/tests/scripts/02_up.sh

@@ -41,7 +41,7 @@ compose() {
   fi
 }
 
-docker container rm -f argus-agent-e2e argus-master-agent-e2e argus-bind-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-e2e argus-agent-env-e2e argus-master-agent-e2e argus-bind-agent-e2e >/dev/null 2>&1 || true
 
 docker network rm tests_default >/dev/null 2>&1 || true
 

src/agent/tests/scripts/03_wait_and_assert_registration.sh

@@ -6,11 +6,14 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 AGENT_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_AGENT_HOSTNAME="host_abc"
 NODE_FILE="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME/node.json"
+ENV_NODE_FILE="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME/node.json"
 
 mkdir -p "$TMP_ROOT"
 
-node_id=""
+primary_node_id=""
+env_node_id=""
 for _ in {1..30}; do
   sleep 2
   response=$(curl -sS "$API_BASE/nodes" || true)
@@ -19,24 +22,49 @@ for _ in {1..30}; do
   fi
   list_file="$TMP_ROOT/nodes_list.json"
   echo "$response" > "$list_file"
-  node_id=$(python3 - "$list_file" <<'PY'
+  readarray -t node_ids < <(python3 - "$list_file" "$AGENT_HOSTNAME" "$ENV_AGENT_HOSTNAME" <<'PY'
 import json, sys
+
 with open(sys.argv[1]) as handle:
     nodes = json.load(handle)
-print(nodes[0]["id"] if nodes else "")
+
+target_primary = sys.argv[2]
+target_env = sys.argv[3]
+
+primary_id = ""
+env_id = ""
+
+for node in nodes:
+    if node.get("name") == target_primary:
+        primary_id = node.get("id", "")
+    if node.get("name") == target_env:
+        env_id = node.get("id", "")
+
+print(primary_id)
+print(env_id)
 PY
-)
-  if [[ -n "$node_id" ]]; then
+  )
+
+  primary_node_id="${node_ids[0]}"
+  env_node_id="${node_ids[1]}"
+
+  if [[ -n "$primary_node_id" && -n "$env_node_id" ]]; then
     break
   fi
  done
 
-if [[ -z "$node_id" ]]; then
-  echo "[ERROR] Agent did not register within timeout" >&2
+if [[ -z "$primary_node_id" ]]; then
+  echo "[ERROR] Primary agent did not register within timeout" >&2
   exit 1
 fi
 
-echo "$node_id" > "$TMP_ROOT/node_id"
+if [[ -z "$env_node_id" ]]; then
+  echo "[ERROR] Env-variable agent did not register within timeout" >&2
+  exit 1
+fi
+
+echo "$primary_node_id" > "$TMP_ROOT/node_id"
+echo "$env_node_id" > "$TMP_ROOT/node_id_host_abc"
 
 if [[ ! -f "$NODE_FILE" ]]; then
   echo "[ERROR] node.json not created at $NODE_FILE" >&2
@@ -50,8 +78,20 @@ with open(sys.argv[1]) as handle:
 assert "id" in node and node["id"], "node.json missing id"
 PY
 
+if [[ ! -f "$ENV_NODE_FILE" ]]; then
+  echo "[ERROR] node.json not created at $ENV_NODE_FILE" >&2
+  exit 1
+fi
+
+python3 - "$ENV_NODE_FILE" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+assert "id" in node and node["id"], "env agent node.json missing id"
+PY
+
 detail_file="$TMP_ROOT/initial_detail.json"
-curl -sS "$API_BASE/nodes/$node_id" -o "$detail_file"
+curl -sS "$API_BASE/nodes/$primary_node_id" -o "$detail_file"
 python3 - "$detail_file" "$TMP_ROOT/initial_ip" <<'PY'
 import json, sys, pathlib
 with open(sys.argv[1]) as handle:
@@ -62,4 +102,5 @@ if not ip:
 pathlib.Path(sys.argv[2]).write_text(ip)
 PY
 
-echo "[INFO] Agent registered with node id $node_id"
+echo "[INFO] Agent registered with node id $primary_node_id"
+echo "[INFO] Env-variable agent registered with node id $env_node_id"

src/agent/tests/scripts/05_verify_agent.sh

@@ -0,0 +1,60 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+REPO_ROOT="$(cd "$TEST_ROOT/.." && pwd)"
+VERIFY_SCRIPT="$REPO_ROOT/scripts/agent_deployment_verify.sh"
+ENV_NODE_ID_FILE="$TEST_ROOT/tmp/node_id_host_abc"
+PRIMARY_CONTAINER="argus-agent-e2e"
+ENV_CONTAINER="argus-agent-env-e2e"
+PRIMARY_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_HOSTNAME="host_abc"
+
+if ! docker ps --format '{{.Names}}' | grep -q "^${PRIMARY_CONTAINER}$"; then
+  echo "[WARN] agent container not running; skip verification"
+  exit 0
+fi
+
+if docker exec -i "$PRIMARY_CONTAINER" bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
+  echo "[INFO] curl/jq already installed in agent container"
+else
+  echo "[INFO] Installing curl/jq in agent container"
+  docker exec -i "$PRIMARY_CONTAINER" bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
+fi
+
+if [[ ! -f "$VERIFY_SCRIPT" ]]; then
+  echo "[ERROR] Verification script missing at $VERIFY_SCRIPT" >&2
+  exit 1
+fi
+
+run_verifier() {
+  local container="$1" hostname="$2"
+
+  if ! docker ps --format '{{.Names}}' | grep -q "^${container}$"; then
+    echo "[WARN] container $container not running; skip"
+    return
+  fi
+
+  if ! docker exec -i "$container" bash -lc 'command -v /usr/local/bin/agent_deployment_verify.sh >/dev/null'; then
+    echo "[ERROR] /usr/local/bin/agent_deployment_verify.sh missing in $container" >&2
+    exit 1
+  fi
+
+  echo "[INFO] Running verification for $hostname in $container"
+  docker exec -i "$container" env VERIFY_HOSTNAME="$hostname" /usr/local/bin/agent_deployment_verify.sh
+}
+
+run_verifier "$PRIMARY_CONTAINER" "$PRIMARY_HOSTNAME"
+
+if docker ps --format '{{.Names}}' | grep -q "^${ENV_CONTAINER}$"; then
+  if docker exec -i "$ENV_CONTAINER" bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
+    echo "[INFO] curl/jq already installed in env agent container"
+  else
+    echo "[INFO] Installing curl/jq in env agent container"
+    docker exec -i "$ENV_CONTAINER" bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
+  fi
+  run_verifier "$ENV_CONTAINER" "$ENV_HOSTNAME"
+else
+  echo "[WARN] env-driven agent container not running; skip secondary verification"
+fi

src/agent/tests/scripts/06_assert_status_on_master.sh

@@ -6,6 +6,8 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 NODE_ID="$(cat "$TMP_ROOT/node_id")"
+ENV_NODE_ID="$(cat "$TMP_ROOT/node_id_host_abc")"
+ENV_HOSTNAME="host_abc"
 NODES_JSON="$TEST_ROOT/private/argus/metric/prometheus/nodes.json"
 
 success=false
@@ -41,13 +43,36 @@ if [[ ! -f "$NODES_JSON" ]]; then
   exit 1
 fi
 
-python3 - "$NODES_JSON" <<'PY'
+python3 - "$NODES_JSON" "$NODE_ID" "$ENV_NODE_ID" <<'PY'
 import json, sys
 with open(sys.argv[1]) as handle:
     nodes = json.load(handle)
-assert len(nodes) == 1, nodes
-entry = nodes[0]
-assert entry["node_id"], entry
+
+expected_primary = sys.argv[2]
+expected_env = sys.argv[3]
+
+ids = {entry.get("node_id") for entry in nodes}
+assert expected_primary in ids, nodes
+assert expected_env in ids, nodes
+assert len(nodes) >= 2, nodes
 PY
 
 echo "[INFO] Master reflects agent health and nodes.json entries"
+
+env_detail_file="$TMP_ROOT/env_agent_detail.json"
+curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_detail_file"
+python3 - "$env_detail_file" "$ENV_HOSTNAME" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+
+expected_name = sys.argv[2]
+
+assert node.get("name") == expected_name, node
+meta = node.get("meta_data", {})
+assert meta.get("env") == "prod", meta
+assert meta.get("user") == "ml", meta
+assert meta.get("instance") == "node-3", meta
+PY
+
+echo "[INFO] Env-variable agent reports expected metadata"

src/agent/tests/scripts/07_restart_agent_and_reregister.sh

@@ -6,10 +6,20 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 NODE_ID="$(cat "$TMP_ROOT/node_id")"
+ENV_NODE_ID_FILE="$TMP_ROOT/node_id_host_abc"
+if [[ ! -f "$ENV_NODE_ID_FILE" ]]; then
+  echo "[ERROR] Env agent node id file missing at $ENV_NODE_ID_FILE" >&2
+  exit 1
+fi
+
+ENV_NODE_ID="$(cat "$ENV_NODE_ID_FILE")"
 AGENT_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_AGENT_HOSTNAME="host_abc"
 NETWORK_NAME="tests_default"
 NEW_AGENT_IP="172.28.0.200"
+NEW_ENV_AGENT_IP="172.28.0.210"
 ENTRYPOINT_SCRIPT="$SCRIPT_DIR/agent_entrypoint.sh"
+VERIFY_SCRIPT="$TEST_ROOT/../scripts/agent_deployment_verify.sh"
 ENV_FILE="$TEST_ROOT/.env"
 
 # 中文提示：重启场景也需要同样的入口脚本，确保 DNS 注册逻辑一致
@@ -18,6 +28,11 @@ if [[ ! -f "$ENTRYPOINT_SCRIPT" ]]; then
   exit 1
 fi
 
+if [[ ! -f "$VERIFY_SCRIPT" ]]; then
+  echo "[ERROR] agent verification script missing at $VERIFY_SCRIPT" >&2
+  exit 1
+fi
+
 if [[ ! -f "$TMP_ROOT/agent_binary_path" ]]; then
   echo "[ERROR] Agent binary path missing; rerun bootstrap" >&2
   exit 1
@@ -74,15 +89,37 @@ if [[ "$prev_ip" != "$initial_ip" ]]; then
   exit 1
 fi
 
+env_before_file="$TMP_ROOT/env_before_restart.json"
+curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_before_file"
+env_prev_last_updated=$(python3 - "$env_before_file" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+print(node.get("last_updated", ""))
+PY
+)
+env_prev_ip=$(python3 - "$env_before_file" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+print(node["meta_data"].get("ip", ""))
+PY
+)
+
 pushd "$TEST_ROOT" >/dev/null
 compose rm -sf agent
+compose rm -sf agent_env
 popd >/dev/null
 
 docker container rm -f argus-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-env-e2e >/dev/null 2>&1 || true
 
 AGENT_DIR="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME"
 HEALTH_DIR="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME/health"
 
+ENV_AGENT_DIR="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME"
+ENV_HEALTH_DIR="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME/health"
+
 # 先以 sleep 方式启动容器，确保我们掌握注册时的网络状态
 if ! docker run -d \
   --name argus-agent-e2e \
@@ -94,6 +131,7 @@ if ! docker run -d \
   -v "$TEST_ROOT/private/argus/etc:/private/argus/etc" \
   -v "$AGENT_BINARY:/usr/local/bin/argus-agent:ro" \
   -v "$ENTRYPOINT_SCRIPT:/usr/local/bin/agent-entrypoint.sh:ro" \
+  -v "$VERIFY_SCRIPT:/usr/local/bin/agent_deployment_verify.sh:ro" \
   -e MASTER_ENDPOINT=http://master.argus.com:3000 \
   -e REPORT_INTERVAL_SECONDS=2 \
   -e ARGUS_BUILD_UID="$AGENT_UID" \
@@ -141,3 +179,76 @@ if [[ "$success" != true ]]; then
 fi
 
 echo "[INFO] Agent restart produced successful re-registration with IP change"
+
+# ---- Restart env-driven agent without metadata environment variables ----
+
+if [[ ! -d "$ENV_AGENT_DIR" ]]; then
+  echo "[ERROR] Env agent data dir missing at $ENV_AGENT_DIR" >&2
+  exit 1
+fi
+
+if [[ ! -d "$ENV_HEALTH_DIR" ]]; then
+  mkdir -p "$ENV_HEALTH_DIR"
+fi
+
+if ! docker run -d \
+  --name argus-agent-env-e2e \
+  --hostname "$ENV_AGENT_HOSTNAME" \
+  --network "$NETWORK_NAME" \
+  --ip "$NEW_ENV_AGENT_IP" \
+  -v "$ENV_AGENT_DIR:/private/argus/agent/$ENV_AGENT_HOSTNAME" \
+  -v "$ENV_HEALTH_DIR:/private/argus/agent/$ENV_AGENT_HOSTNAME/health" \
+  -v "$TEST_ROOT/private/argus/etc:/private/argus/etc" \
+  -v "$AGENT_BINARY:/usr/local/bin/argus-agent:ro" \
+  -v "$ENTRYPOINT_SCRIPT:/usr/local/bin/agent-entrypoint.sh:ro" \
+  -v "$VERIFY_SCRIPT:/usr/local/bin/agent_deployment_verify.sh:ro" \
+  -e MASTER_ENDPOINT=http://master.argus.com:3000 \
+  -e REPORT_INTERVAL_SECONDS=2 \
+  -e ARGUS_BUILD_UID="$AGENT_UID" \
+  -e ARGUS_BUILD_GID="$AGENT_GID" \
+  --entrypoint /usr/local/bin/agent-entrypoint.sh \
+  ubuntu:22.04 >/dev/null; then
+  echo "[ERROR] Failed to start env-driven agent container without metadata env" >&2
+  exit 1
+fi
+
+env_success=false
+env_detail_file="$TMP_ROOT/env_post_restart.json"
+for _ in {1..20}; do
+  sleep 3
+  if ! curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_detail_file"; then
+    continue
+  fi
+  if python3 - "$env_detail_file" "$env_prev_last_updated" "$ENV_NODE_ID" "$env_prev_ip" "$NEW_ENV_AGENT_IP" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+prev_last_updated = sys.argv[2]
+expected_id = sys.argv[3]
+old_ip = sys.argv[4]
+expected_ip = sys.argv[5]
+last_updated = node.get("last_updated")
+current_ip = node["meta_data"].get("ip")
+meta = node.get("meta_data", {})
+assert node["id"] == expected_id
+if current_ip != expected_ip:
+    raise SystemExit(1)
+if current_ip == old_ip:
+    raise SystemExit(1)
+if not last_updated or last_updated == prev_last_updated:
+    raise SystemExit(1)
+if meta.get("env") != "prod" or meta.get("user") != "ml" or meta.get("instance") != "node-3":
+    raise SystemExit(1)
+PY
+  then
+    env_success=true
+    break
+  fi
+done
+
+if [[ "$env_success" != true ]]; then
+  echo "[ERROR] Env-driven agent did not reuse persisted metadata after restart" >&2
+  exit 1
+fi
+
+echo "[INFO] Env-driven agent restart succeeded with persisted metadata"

src/agent/tests/scripts/08_down.sh

@@ -13,7 +13,7 @@ compose() {
   fi
 }
 
-docker container rm -f argus-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-e2e argus-agent-env-e2e >/dev/null 2>&1 || true
 
 pushd "$TEST_ROOT" >/dev/null
 compose down --remove-orphans

src/agent/tests/scripts/08_verify_agent.sh

@@ -1,26 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
-VERIFY_SCRIPT="$(cd "$TEST_ROOT/.." && pwd)/scripts/agent_deployment_verify.sh"
-
-if ! docker ps --format '{{.Names}}' | grep -q '^argus-agent-e2e$'; then
-  echo "[WARN] agent container not running; skip verification"
-  exit 0
-fi
-
-if docker exec -i argus-agent-e2e bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
-  echo "[INFO] curl/jq already installed in agent container"
-else
-  echo "[INFO] Installing curl/jq in agent container"
-  docker exec -i argus-agent-e2e bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
-fi
-
-if docker exec -i argus-agent-e2e bash -lc 'command -v /usr/local/bin/agent_deployment_verify.sh >/dev/null'; then
-  docker exec -i argus-agent-e2e /usr/local/bin/agent_deployment_verify.sh
-elif [[ -x "$VERIFY_SCRIPT" ]]; then
-  docker exec -i argus-agent-e2e "$VERIFY_SCRIPT"
-else
-  echo "[WARN] agent_deployment_verify.sh not found"
-fi

src/agent/tests/test_config_metadata.py

@@ -0,0 +1,151 @@
+from __future__ import annotations
+
+import os
+import unittest
+from contextlib import contextmanager
+from unittest.mock import patch
+
+from app.config import AgentConfig, load_config
+
+
+@contextmanager
+def temp_env(**overrides: str | None):
+    originals: dict[str, str | None] = {}
+    try:
+        for key, value in overrides.items():
+            originals[key] = os.environ.get(key)
+            if value is None:
+                os.environ.pop(key, None)
+            else:
+                os.environ[key] = value
+        yield
+    finally:
+        for key, original in originals.items():
+            if original is None:
+                os.environ.pop(key, None)
+            else:
+                os.environ[key] = original
+
+
+class LoadConfigMetadataTests(unittest.TestCase):
+    @patch("app.config.Path.mkdir")
+    def test_metadata_from_environment_variables(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="dev-user-one-pod",
+            AGENT_ENV="prod",
+            AGENT_USER="ops",
+            AGENT_INSTANCE="node-1",
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "prod")
+        self.assertEqual(config.user, "ops")
+        self.assertEqual(config.instance, "node-1")
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_metadata_falls_back_to_hostname(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="qa-team-abc-pod-2",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "qa")
+        self.assertEqual(config.user, "team")
+        self.assertEqual(config.instance, "abc")
+        mock_mkdir.assert_called()
+
+    @patch("app.config._load_metadata_from_state", return_value=("prod", "ops", "node-1"))
+    @patch("app.config.Path.mkdir")
+    def test_metadata_from_node_state(self, mock_mkdir, mock_state):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="host_abc",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "prod")
+        self.assertEqual(config.user, "ops")
+        self.assertEqual(config.instance, "node-1")
+        mock_state.assert_called_once()
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_partial_environment_variables_fallback(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="stage-ml-001-node",
+            AGENT_ENV="prod",
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "stage")
+        self.assertEqual(config.user, "ml")
+        self.assertEqual(config.instance, "001")
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_invalid_hostname_raises_error(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="invalidhostname",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            with self.assertRaises(ValueError):
+                load_config()
+
+        mock_mkdir.assert_not_called()
+
+
+class CollectMetadataTests(unittest.TestCase):
+    @patch("app.collector._detect_ip_address", return_value="127.0.0.1")
+    @patch("app.collector._detect_gpu_count", return_value=0)
+    @patch("app.collector._detect_memory_bytes", return_value=1024)
+    @patch("app.collector._detect_cpu_count", return_value=8)
+    def test_collect_metadata_uses_config_fields(
+        self,
+        mock_cpu,
+        mock_memory,
+        mock_gpu,
+        mock_ip,
+    ):
+        config = AgentConfig(
+            hostname="dev-user-001-pod",
+            environment="prod",
+            user="ops",
+            instance="node-1",
+            node_file="/tmp/node.json",
+            version="1.0.0",
+            master_endpoint="http://master.local",
+            report_interval_seconds=60,
+            health_dir="/tmp/health",
+        )
+
+        from app.collector import collect_metadata
+
+        metadata = collect_metadata(config)
+
+        self.assertEqual(metadata["env"], "prod")
+        self.assertEqual(metadata["user"], "ops")
+        self.assertEqual(metadata["instance"], "node-1")
+        self.assertEqual(metadata["hostname"], "dev-user-001-pod")
+        self.assertEqual(metadata["ip"], "127.0.0.1")
+        self.assertEqual(metadata["cpu_number"], 8)
+        self.assertEqual(metadata["memory_in_bytes"], 1024)
+        self.assertEqual(metadata["gpu_number"], 0)
+
+
+if __name__ == "__main__":
+    unittest.main()

src/alert/README.md

@@ -0,0 +1,31 @@
+# Alertmanager
+
+## 构建
+1. 首先设置构建和部署的环境变量, 在项目根目录下执行：
+```bash
+cp src/alert/tests/.env.example src/alert/tests/.env
+```
+
+然后找到复制出来的.env文件，修改环境变量。
+
+2. 使用脚本构建，在项目根目录下执行：
+
+```bash
+bash src/alert/alertmanager/build/build.sh
+```
+
+构建成功后，会在项目根目录下生成argus-alertmanager-latest.tar
+
+## 部署
+
+提供docker-compose部署。在src/alert/tests目录下
+```bash
+docker-compose up -d
+```
+
+## 动态配置
+配置文件放在`/private/argus/alert/alertmanager/alertmanager.yml`下，修改alertmanager.yml后，调用`http://alertmanager.alert.argus.com:9093/-/reload`接口(POST)可以重新加载配置.
+
+```bash
+curl -X POST http://localhost:9093/-/reload
+```

src/alert/alertmanager/build/Dockerfile

@@ -0,0 +1,96 @@
+# 基于 Ubuntu 24.04
+FROM ubuntu:24.04
+
+# 切换到 root 用户
+USER root
+
+# 安装必要依赖
+RUN apt-get update && \
+    apt-get install -y wget supervisor net-tools inetutils-ping vim ca-certificates passwd && \
+    apt-get clean && rm -rf /var/lib/apt/lists/*
+
+# 设置 Alertmanager 版本（与本地离线包保持一致）
+ARG ALERTMANAGER_VERSION=0.28.1
+
+# 使用仓库内预置的离线包构建（无需联网）
+COPY src/alert/alertmanager/build/alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz /tmp/
+RUN tar xvf /tmp/alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz -C /tmp && \
+    mv /tmp/alertmanager-${ALERTMANAGER_VERSION}.linux-amd64 /usr/local/alertmanager && \
+    rm -f /tmp/alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz
+
+ENV ALERTMANAGER_BASE_PATH=/private/argus/alert/alertmanager
+
+ARG ARGUS_BUILD_UID=2133
+ARG ARGUS_BUILD_GID=2015
+ENV ARGUS_BUILD_UID=${ARGUS_BUILD_UID}
+ENV ARGUS_BUILD_GID=${ARGUS_BUILD_GID}
+
+RUN mkdir -p /usr/share/alertmanager && \
+    mkdir -p ${ALERTMANAGER_BASE_PATH} && \
+    mkdir -p /private/argus/etc && \
+    rm -rf /alertmanager && \
+    ln -s ${ALERTMANAGER_BASE_PATH} /alertmanager
+
+# 创建 alertmanager 用户（可自定义 UID/GID）
+# 创建 alertmanager 用户组
+RUN set -eux; \
+    # 确保目标 GID 存在；若已被占用，直接使用该 GID（组名不限）\
+    if ! getent group "${ARGUS_BUILD_GID}" >/dev/null; then \
+        groupadd -g "${ARGUS_BUILD_GID}" alertmanager || true; \
+    fi; \
+    # 确保存在 alertmanager 用户；若 UID 已被占用，跳过并继续使用现有 UID 的用户
+    if ! id alertmanager >/dev/null 2>&1; then \
+        if getent passwd "${ARGUS_BUILD_UID}" >/dev/null; then \
+            # UID 已占用，则创建同名用户但不指定 UID（避免冲突），仅保证 user 存在
+            useradd -M -s /usr/sbin/nologin -g "${ARGUS_BUILD_GID}" alertmanager || true; \
+        else \
+            useradd -M -s /usr/sbin/nologin -u "${ARGUS_BUILD_UID}" -g "${ARGUS_BUILD_GID}" alertmanager || true; \
+        fi; \
+    else \
+        usermod -g "${ARGUS_BUILD_GID}" alertmanager || true; \
+    fi
+
+RUN chown -R "${ARGUS_BUILD_UID}:${ARGUS_BUILD_GID}" /usr/share/alertmanager /alertmanager ${ALERTMANAGER_BASE_PATH} /private/argus/etc /usr/local/bin || true
+
+# 配置内网 apt 源 (如果指定了内网选项)
+RUN if [ "$USE_INTRANET" = "true" ]; then \
+        echo "Configuring intranet apt sources..." && \
+        cp /etc/apt/sources.list /etc/apt/sources.list.bak && \
+        echo "deb [trusted=yes] http://10.68.64.1/ubuntu2204/ jammy main" > /etc/apt/sources.list && \
+        echo 'Acquire::https::Verify-Peer "false";' > /etc/apt/apt.conf.d/99disable-ssl-check && \
+        echo 'Acquire::https::Verify-Host "false";' >> /etc/apt/apt.conf.d/99disable-ssl-check; \
+    fi
+
+
+# 配置部署时使用的 apt 源
+RUN if [ "$USE_INTRANET" = "true" ]; then \
+    echo "deb [trusted=yes] https://10.92.132.52/mirrors/ubuntu2204/ jammy main" > /etc/apt/sources.list; \
+    fi
+
+# 创建 supervisor 日志目录
+RUN mkdir -p /var/log/supervisor
+
+# 复制 supervisor 配置文件
+COPY src/alert/alertmanager/build/supervisord.conf /etc/supervisor/conf.d/supervisord.conf
+
+# 复制启动脚本
+COPY src/alert/alertmanager/build/start-am-supervised.sh /usr/local/bin/start-am-supervised.sh
+RUN chmod +x /usr/local/bin/start-am-supervised.sh
+
+# 复制 Alertmanager 配置文件
+COPY src/alert/alertmanager/build/alertmanager.yml /etc/alertmanager/alertmanager.yml
+RUN chmod +x /etc/alertmanager/alertmanager.yml
+# COPY src/alert/alertmanager/build/alertmanager.yml ${ALERTMANAGER_BASE_PATH}/alertmanager.yml
+
+# 复制 DNS 监控脚本
+COPY src/alert/alertmanager/build/dns-monitor.sh /usr/local/bin/dns-monitor.sh
+RUN chmod +x /usr/local/bin/dns-monitor.sh
+
+# 保持 root 用户，由 supervisor 控制 user 切换
+USER root
+
+# 暴露端口（Alertmanager 默认端口 9093）
+EXPOSE 9093
+
+# 使用 supervisor 作为入口点
+CMD ["/usr/bin/supervisord", "-c", "/etc/supervisor/conf.d/supervisord.conf"]

src/alert/alertmanager/build/alertmanager.yml

@@ -0,0 +1,19 @@
+global:
+  resolve_timeout: 5m
+
+route:
+  group_by: ['alertname', 'instance']   # 分组：相同 alertname + instance 的告警合并
+  group_wait: 30s        # 第一个告警后，等 30s 看是否有同组告警一起发
+  group_interval: 5m     # 同组告警变化后，至少 5 分钟再发一次
+  repeat_interval: 3h    # 相同告警，3 小时重复提醒一次
+  receiver: 'null'
+
+receivers:
+  - name: 'null'
+
+inhibit_rules:
+  - source_match:
+      severity: 'critical'     # critical 告警存在时
+    target_match:
+      severity: 'warning'      # 抑制相同 instance 的 warning 告警
+    equal: ['instance']

src/alert/alertmanager/build/build.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -euo pipefail
+docker pull ubuntu:24.04
+
+source src/alert/tests/.env
+
+docker build \
+  --build-arg ARGUS_BUILD_UID=${ARGUS_BUILD_UID} \
+  --build-arg ARGUS_BUILD_GID=${ARGUS_BUILD_GID} \
+  -f src/alert/alertmanager/build/Dockerfile \
+  -t argus-alertmanager:latest .
+
+docker save -o argus-alertmanager-latest.tar argus-alertmanager:latest

src/alert/alertmanager/build/dns-monitor.sh

@@ -0,0 +1,68 @@
+#!/bin/bash
+
+# DNS监控脚本 - 每10秒检查dns.conf是否有变化
+# 如果有变化则执行update-dns.sh脚本
+
+DNS_CONF="/private/argus/etc/dns.conf"
+DNS_BACKUP="/tmp/dns.conf.backup"
+UPDATE_SCRIPT="/private/argus/etc/update-dns.sh"
+LOG_FILE="/var/log/supervisor/dns-monitor.log"
+
+# 确保日志文件存在
+touch "$LOG_FILE"
+
+log_message() {
+    echo "$(date '+%Y-%m-%d %H:%M:%S') [DNS-Monitor] $1" >> "$LOG_FILE"
+}
+
+log_message "DNS监控脚本启动"
+
+while true; do
+    if [ -f "$DNS_CONF" ]; then
+        if [ -f "$DNS_BACKUP" ]; then
+            # 比较文件内容
+            if ! cmp -s "$DNS_CONF" "$DNS_BACKUP"; then
+                log_message "检测到DNS配置变化"
+
+                # 更新备份文件
+                cp "$DNS_CONF" "$DNS_BACKUP"
+
+                # 执行更新脚本
+                if [ -x "$UPDATE_SCRIPT" ]; then
+                    log_message "执行DNS更新脚本: $UPDATE_SCRIPT"
+                    "$UPDATE_SCRIPT" >> "$LOG_FILE" 2>&1
+                    if [ $? -eq 0 ]; then
+                        log_message "DNS更新脚本执行成功"
+                    else
+                        log_message "DNS更新脚本执行失败"
+                    fi
+                else
+                    log_message "警告: 更新脚本不存在或不可执行: $UPDATE_SCRIPT"
+                fi
+            fi
+        else
+
+            # 第一次检测到配置文件，执行更新脚本
+            if [ -x "$UPDATE_SCRIPT" ]; then
+                log_message "执行DNS更新脚本: $UPDATE_SCRIPT"
+                "$UPDATE_SCRIPT" >> "$LOG_FILE" 2>&1
+                if [ $? -eq 0 ]; then
+                    log_message "DNS更新脚本执行成功"
+
+		    # 第一次运行，创建备份并执行更新
+		    cp "$DNS_CONF" "$DNS_BACKUP"
+		    log_message "创建DNS配置备份文件"
+
+                else
+                    log_message "DNS更新脚本执行失败"
+                fi
+            else
+                log_message "警告: 更新脚本不存在或不可执行: $UPDATE_SCRIPT"
+            fi
+        fi
+    else
+        log_message "警告: DNS配置文件不存在: $DNS_CONF"
+    fi
+
+    sleep 10
+done

src/alert/alertmanager/build/fetch-dist.sh

@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# 下载 Alertmanager 离线安装包到本目录，用于 Docker 构建时 COPY
+# 用法：
+#   ./fetch-dist.sh [version]
+# 示例：
+#   ./fetch-dist.sh 0.28.1
+
+VER="${1:-0.28.1}"
+OUT="alertmanager-${VER}.linux-amd64.tar.gz"
+URL="https://github.com/prometheus/alertmanager/releases/download/v${VER}/${OUT}"
+
+if [[ -f "$OUT" ]]; then
+  echo "[INFO] $OUT already exists, skip download"
+  exit 0
+fi
+
+echo "[INFO] Downloading $URL"
+curl -fL --retry 3 --connect-timeout 10 -o "$OUT" "$URL"
+echo "[OK] Saved to $(pwd)/$OUT"
+

src/alert/alertmanager/build/start-am-supervised.sh

@@ -0,0 +1,25 @@
+#!/bin/bash
+set -euo pipefail
+
+echo "[INFO] Starting Alertmanager under supervisor..."
+
+ALERTMANAGER_BASE_PATH=${ALERTMANAGER_BASE_PATH:-/private/argus/alert/alertmanager}
+
+echo "[INFO] Alertmanager base path: ${ALERTMANAGER_BASE_PATH}"
+
+# 使用容器内的 /etc/alertmanager/alertmanager.yml 作为配置文件，避免写入挂载卷导致的权限问题
+echo "[INFO] Using /etc/alertmanager/alertmanager.yml as configuration"
+
+
+# 记录容器 IP 地址
+DOMAIN=alertmanager.alert.argus.com
+IP=$(ifconfig | grep -A 1 eth0 | grep inet | awk '{print $2}')
+echo "current IP: ${IP}"
+echo "${IP}" > /private/argus/etc/${DOMAIN}
+chmod 755 /private/argus/etc/${DOMAIN}
+
+
+echo "[INFO] Starting Alertmanager process..."
+
+# 启动 Alertmanager 主进程
+exec /usr/local/alertmanager/alertmanager --config.file=/etc/alertmanager/alertmanager.yml --storage.path=/alertmanager --cluster.listen-address=""

src/alert/alertmanager/build/supervisord.conf

@@ -0,0 +1,39 @@
+[supervisord]
+nodaemon=true
+logfile=/var/log/supervisor/supervisord.log
+pidfile=/var/run/supervisord.pid
+user=root
+
+[program:alertmanager]
+command=/usr/local/bin/start-am-supervised.sh
+user=ubuntu
+stdout_logfile=/var/log/supervisor/alertmanager.log
+stderr_logfile=/var/log/supervisor/alertmanager_error.log
+autorestart=true
+startretries=3
+startsecs=10
+stopwaitsecs=20
+killasgroup=true
+stopasgroup=true
+
+[program:dns-monitor]
+command=/usr/local/bin/dns-monitor.sh
+user=root
+stdout_logfile=/var/log/supervisor/dns-monitor.log
+stderr_logfile=/var/log/supervisor/dns-monitor_error.log
+autorestart=true
+startretries=3
+startsecs=5
+stopwaitsecs=10
+killasgroup=true
+stopasgroup=true
+
+[unix_http_server]
+file=/var/run/supervisor.sock
+chmod=0700
+
+[supervisorctl]
+serverurl=unix:///var/run/supervisor.sock
+
+[rpcinterface:supervisor]
+supervisor.rpcinterface_factory = supervisor.rpcinterface:make_main_rpcinterface

src/alert/alertmanager/config/rule_files/README.md

@@ -0,0 +1,60 @@
+# 告警配置
+
+> 参考：[自定义Prometheus告警规则](https://yunlzheng.gitbook.io/prometheus-book/parti-prometheus-ji-chu/alert/prometheus-alert-rule)
+
+在Prometheus中配置告警的有两个步骤：
+
+1. 写告警规则文件（rules文件）
+2. 在promethues.yml里加载规则，并配置Alertmanager
+
+## 1. 编写告警规则文件
+告警规则如下：
+```yml
+groups:
+  - name: example-rules
+    interval: 30s  # 每30秒评估一次
+    rules:
+      - alert: InstanceDown
+        expr: up == 0
+        for: 1m
+        labels:
+          severity: critical
+        annotations:
+          summary: "实例 {{ $labels.instance }} 已宕机"
+          description: "{{ $labels.instance }} 在 {{ $labels.job }} 中无响应超过 1 分钟。"
+
+      - alert: HighCpuUsage
+        expr: 100 - (avg by (instance) (irate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 80
+        for: 5m
+        labels:
+          severity: warning
+        annotations:
+          summary: "CPU 使用率过高"
+          description: "实例 {{ $labels.instance }} CPU 使用率超过 80% 持续 5 分钟。"
+```
+
+其中：
+
+- `alert`：告警规则的名称。
+- `expr`：基于PromQL表达式告警触发条件，用于计算是否有时间序列满足该条件。
+- `for`：评估等待时间，可选参数。用于表示只有当触发条件持续一段时间后才发送告警。在等待期间新产生告警的状态为pending。
+- `labels`：自定义标签，允许用户指定要附加到告警上的一组附加标签，可以在Alertmanager中做路由和分组。
+- `annotations`：用于指定一组附加信息，比如用于描述告警详细信息的文字等，annotations的内容在告警产生时会一同作为参数发送到Alertmanager。可以提供告警摘要和详细信息。
+
+## 2. promothues.yml里引用
+在prometheus.yml中加上`rule_files`和`alerting`:
+
+```yml
+global:
+  [ evaluation_interval: <duration> | default = 1m ]
+
+rule_files:
+  [ - <filepath_glob> ... ]
+
+alerting:
+  alertmanagers:
+    - static_configs:
+        - targets:
+            - "alertmanager.alert.argus.com:9093"   # Alertmanager 地址
+
+```

src/alert/alertmanager/config/rule_files/example_rules.yml

@@ -0,0 +1,37 @@
+groups:
+  - name: example-rules
+    interval: 30s  # 每30秒评估一次
+    rules:
+      - alert: InstanceDown
+        expr: up == 0
+        for: 1m
+        labels:
+          severity: critical
+        annotations:
+          summary: "实例 {{ $labels.instance }} 已宕机"
+          description: "{{ $labels.instance }} 在 {{ $labels.job }} 中无响应超过 1 分钟。"
+
+      - alert: HighCpuUsage
+        expr: 100 - (avg by (instance) (irate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 80
+        for: 5m
+        labels:
+          severity: warning
+        annotations:
+          summary: "CPU 使用率过高"
+          description: "实例 {{ $labels.instance }} CPU 使用率超过 80% 持续 5 分钟。"
+      - alert: HighMemoryUsage
+        expr: (node_memory_MemTotal_bytes - node_memory_MemAvailable_bytes) / node_memory_MemTotal_bytes * 100 > 80
+        for: 5m
+        labels:
+          severity: warning
+        annotations:
+          summary: "内存使用率过高"
+          description: "实例 {{ $labels.instance }} 内存使用率超过 80% 持续 5 分钟。"
+      - alert: DiskSpaceLow
+        expr: (node_filesystem_size_bytes{fstype!~"tmpfs|overlay"} - node_filesystem_free_bytes{fstype!~"tmpfs|overlay"}) / node_filesystem_size_bytes{fstype!~"tmpfs|overlay"} * 100 > 90
+        for: 10m
+        labels:
+          severity: warning
+        annotations:
+          summary: "磁盘空间不足"
+          description: "实例 {{ $labels.instance }} 磁盘空间不足超过 90% 持续 10 分钟。"

src/alert/tests/.env

@@ -0,0 +1,5 @@
+DATA_ROOT=/home/argus/tmp/private/argus
+ARGUS_BUILD_UID=1048
+ARGUS_BUILD_GID=1048
+
+USE_INTRANET=false

src/alert/tests/.env.example

@@ -0,0 +1,5 @@
+DATA_ROOT=/home/argus/tmp/private/argus
+ARGUS_UID=1048
+ARGUS_GID=1048
+
+USE_INTRANET=false

src/alert/tests/data/alertmanager/alertmanager.yml

@@ -0,0 +1,19 @@
+global:
+  resolve_timeout: 5m
+
+route:
+  group_by: ['alertname', 'instance']   # 分组：相同 alertname + instance 的告警合并
+  group_wait: 30s        # 第一个告警后，等 30s 看是否有同组告警一起发
+  group_interval: 5m     # 同组告警变化后，至少 5 分钟再发一次
+  repeat_interval: 3h    # 相同告警，3 小时重复提醒一次
+  receiver: 'null'
+
+receivers:
+  - name: 'null'
+
+inhibit_rules:
+  - source_match:
+      severity: 'critical'     # critical 告警存在时
+    target_match:
+      severity: 'warning'      # 抑制相同 instance 的 warning 告警
+    equal: ['instance']

src/alert/tests/data/etc/alertmanager.alert.argus.com

@@ -0,0 +1 @@
+172.18.0.2

src/alert/tests/docker-compose.yml

@@ -0,0 +1,37 @@
+services:
+  alertmanager:
+    build:
+      context: ../../../
+      dockerfile: src/alert/alertmanager/build/Dockerfile
+      args:
+        ARGUS_BUILD_UID: ${ARGUS_BUILD_UID:-2133}
+        ARGUS_BUILD_GID: ${ARGUS_BUILD_GID:-2015}
+        USE_INTRANET: ${USE_INTRANET:-false}
+    image: argus-alertmanager:latest
+    container_name: argus-alertmanager
+    environment:
+      - ALERTMANAGER_BASE_PATH=/private/argus/alert/alertmanager
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+    ports:
+      - "${ARGUS_PORT:-9093}:9093"
+    volumes:
+      - ${DATA_ROOT:-./data}/alert/alertmanager:/private/argus/alert/alertmanager
+      - ${DATA_ROOT:-./data}/etc:/private/argus/etc
+    networks:
+      - argus-debug-net
+    restart: unless-stopped
+    logging:
+      driver: "json-file"
+      options:
+        max-size: "10m"
+        max-file: "3"
+
+networks:
+  argus-debug-net:
+    driver: bridge
+    name: argus-debug-net
+
+volumes:
+  alertmanager_data:
+    driver: local

src/alert/tests/scripts/01_bootstrap.sh

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+set -euo pipefail
+root="$(cd "$(dirname "${BASH_SOURCE[0]}")/../" && pwd)"
+project_root="$(cd "$root/../../.." && pwd)"
+
+source "$project_root/scripts/common/build_user.sh"
+load_build_user
+
+# 创建新的private目录结构 (基于argus目录结构)
+echo "[INFO] Creating private directory structure for supervisor-based containers..."
+mkdir -p "$root/private/argus/alert/alertmanager"
+mkdir -p "$root/private/argus/etc/"
+
+# 设置数据目录权限
+echo "[INFO] Setting permissions for data directories..."
+chown -R "${ARGUS_BUILD_UID}:${ARGUS_BUILD_GID}" "$root/private/argus/alert/alertmanager" 2>/dev/null || true
+chown -R "${ARGUS_BUILD_UID}:${ARGUS_BUILD_GID}" "$root/private/argus/etc" 2>/dev/null || true
+
+echo "[INFO] Supervisor-based containers will manage their own scripts and configurations"

src/alert/tests/scripts/02_up.sh

@@ -0,0 +1,10 @@
+#!/usr/bin/env bash
+set -euo pipefail
+cd "$(dirname "$0")/.."
+compose_cmd="docker compose"
+if ! $compose_cmd version >/dev/null 2>&1; then
+  if command -v docker-compose >/dev/null 2>&1; then compose_cmd="docker-compose"; else
+    echo "需要 Docker Compose，请安装后重试" >&2; exit 1; fi
+fi
+$compose_cmd -p alert-mvp up -d --remove-orphans
+echo "[OK] 服务已启动：Alertmanager http://localhost:9093"

src/alert/tests/scripts/03_alertmanager_add_alert.sh

@@ -0,0 +1,106 @@
+#!/bin/bash
+set -euo pipefail
+
+# ==========================================================
+# Alertmanager 测试脚本
+# ==========================================================
+
+ALERTMANAGER_URL="http://localhost:9093"
+TEST_ALERT_NAME_CRITICAL="NodeDown"
+TEST_ALERT_NAME_WARNING="HighCPU"
+TMP_LOG="/tmp/test-alertmanager.log"
+
+# 等待参数
+am_wait_attempts=30
+am_wait_interval=2
+
+GREEN="\033[1;32m"
+RED="\033[1;31m"
+YELLOW="\033[1;33m"
+RESET="\033[0m"
+
+# ==========================================================
+# 函数定义
+# ==========================================================
+
+wait_for_alertmanager() {
+  local attempt=1
+  echo "[INFO] 等待 Alertmanager 启动中..."
+  while (( attempt <= am_wait_attempts )); do
+    if curl -fsS "${ALERTMANAGER_URL}/api/v2/status" >/dev/null 2>&1; then
+      echo -e "${GREEN}[OK] Alertmanager 已就绪 (attempt=${attempt}/${am_wait_attempts})${RESET}"
+      return 0
+    fi
+    echo "[..] Alertmanager 尚未就绪 (${attempt}/${am_wait_attempts})"
+    sleep "${am_wait_interval}"
+    (( attempt++ ))
+  done
+  echo -e "${RED}[ERROR] Alertmanager 在 ${am_wait_attempts} 次尝试后仍未就绪${RESET}"
+  return 1
+}
+
+log_step() {
+  echo -e "${YELLOW}==== $1 ====${RESET}"
+}
+
+# ==========================================================
+# 主流程
+# ==========================================================
+
+log_step "测试 Alertmanager 开始"
+echo "[INFO] Alertmanager 地址: $ALERTMANAGER_URL"
+
+# Step 1: 等待 Alertmanager 启动
+wait_for_alertmanager
+
+# Step 2: 触发一个critical测试告警
+echo "[INFO] 发送critical测试告警..."
+curl -fsS -X POST "${ALERTMANAGER_URL}/api/v2/alerts" \
+  -H "Content-Type: application/json" \
+  -d '[
+    {
+      "labels": {
+        "alertname": "'"${TEST_ALERT_NAME_CRITICAL}"'",
+        "instance": "node-1",
+        "severity": "critical"
+      },
+      "annotations": {
+        "summary": "节点 node-1 宕机"
+      }
+    }
+  ]' \
+  -o "$TMP_LOG"
+
+if [ $? -eq 0 ]; then
+  echo -e "${GREEN}[OK] 已成功发送critical测试告警${RESET}"
+else
+  echo -e "${RED}[ERROR] critical告警发送失败！${RESET}"
+  cat "$TMP_LOG"
+  exit 1
+fi
+
+# Step 3: 触发一个warning测试告警
+echo "[INFO] 发送warning测试告警..."
+curl -fsS -X POST "${ALERTMANAGER_URL}/api/v2/alerts" \
+  -H "Content-Type: application/json" \
+  -d '[
+    {
+      "labels": {
+        "alertname": "'"${TEST_ALERT_NAME_WARNING}"'",
+        "instance": "node-1",
+        "severity": "warning"
+      },
+      "annotations": {
+        "summary": "节点 node-1 CPU 使用率过高"
+      }
+    }
+  ]' \
+  -o "$TMP_LOG"
+
+if [ $? -eq 0 ]; then
+  echo -e "${GREEN}[OK] 已成功发送warning测试告警${RESET}"
+else
+  echo -e "${RED}[ERROR] warning告警发送失败！${RESET}"
+  cat "$TMP_LOG"
+  exit 1
+fi

src/alert/tests/scripts/04_query_alerts.sh

@@ -0,0 +1,71 @@
+#!/bin/bash
+set -euo pipefail
+
+# ==========================================================
+# Alertmanager 测试脚本（含启动等待）
+# ==========================================================
+
+ALERTMANAGER_URL="http://localhost:9093"
+TEST_ALERT_NAME_CRITICAL="NodeDown"
+TEST_ALERT_NAME_WARNING="HighCPU"
+TMP_LOG="/tmp/test-alertmanager.log"
+
+# 等待参数
+am_wait_attempts=30
+am_wait_interval=2
+
+GREEN="\033[1;32m"
+RED="\033[1;31m"
+YELLOW="\033[1;33m"
+RESET="\033[0m"
+
+# ==========================================================
+# 函数定义
+# ==========================================================
+
+wait_for_alertmanager() {
+  local attempt=1
+  echo "[INFO] 等待 Alertmanager 启动中..."
+  while (( attempt <= am_wait_attempts )); do
+    if curl -fsS "${ALERTMANAGER_URL}/api/v2/status" >/dev/null 2>&1; then
+      echo -e "${GREEN}[OK] Alertmanager 已就绪 (attempt=${attempt}/${am_wait_attempts})${RESET}"
+      return 0
+    fi
+    echo "[..] Alertmanager 尚未就绪 (${attempt}/${am_wait_attempts})"
+    sleep "${am_wait_interval}"
+    (( attempt++ ))
+  done
+  echo -e "${RED}[ERROR] Alertmanager 在 ${am_wait_attempts} 次尝试后仍未就绪${RESET}"
+  return 1
+}
+
+log_step() {
+  echo -e "${YELLOW}==== $1 ====${RESET}"
+}
+
+# ==========================================================
+# 主流程
+# ==========================================================
+
+log_step "查询 Alertmanager 当前告警列表开始"
+echo "[INFO] Alertmanager 地址: $ALERTMANAGER_URL"
+
+# Step 1: 等待 Alertmanager 启动
+wait_for_alertmanager
+
+# Step 2: 查询当前告警列表
+echo "[INFO] 查询当前告警..."
+sleep 1
+curl -fsS "${ALERTMANAGER_URL}/api/v2/alerts" | jq '.' || {
+  echo -e "${RED}[WARN] 无法解析返回 JSON，请检查 jq 是否安装${RESET}"
+  curl -s "${ALERTMANAGER_URL}/api/v2/alerts"
+}
+
+# Step 3: 检查告警是否包含 NodeDown
+if curl -fsS "${ALERTMANAGER_URL}/api/v2/alerts" | grep -q "${TEST_ALERT_NAME_CRITICAL}"; then
+  echo -e "${GREEN}✅ 测试通过：Alertmanager 已成功接收告警 ${TEST_ALERT_NAME_CRITICAL}${RESET}"
+else
+  echo -e "${RED}❌ 测试失败：未检测到告警 ${TEST_ALERT_NAME_CRITICAL}${RESET}"
+fi
+
+log_step "测试结束"

src/alert/tests/scripts/05_down.sh

@@ -0,0 +1,21 @@
+#!/usr/bin/env bash
+set -euo pipefail
+cd "$(dirname "$0")/.."
+compose_cmd="docker compose"
+if ! $compose_cmd version >/dev/null 2>&1; then
+  if command -v docker-compose >/dev/null 2>&1; then compose_cmd="docker-compose"; else
+    echo "需要 Docker Compose，请安装后重试" >&2; exit 1; fi
+fi
+$compose_cmd -p alert-mvp down
+echo "[OK] 已停止所有容器"
+
+# 清理private目录内容
+echo "[INFO] 清理private目录内容..."
+cd "$(dirname "$0")/.."
+if [ -d "private" ]; then
+    # 删除private目录及其所有内容
+    rm -rf private
+    echo "[OK] 已清理private目录"
+else
+    echo "[INFO] private目录不存在，无需清理"
+fi

src/alert/tests/scripts/e2e_test.sh

@@ -0,0 +1,105 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+echo "======================================="
+echo "ARGUS Alert System End-to-End Test"
+echo "======================================="
+echo ""
+
+# 记录测试开始时间
+test_start_time=$(date +%s)
+
+# 函数：等待服务就绪
+wait_for_services() {
+    echo "[INFO] Waiting for all services to be ready..."
+    local max_attempts=${SERVICE_WAIT_ATTEMPTS:-120}
+    local attempt=1
+
+    while [ $attempt -le $max_attempts ]; do
+        if curl -fs http://localhost:9093/api/v2/status >/dev/null 2>&1; then
+            echo "[OK] All services are ready!"
+            return 0
+        fi
+        echo "    Waiting for services... ($attempt/$max_attempts)"
+        sleep 5
+        ((attempt++))
+    done
+
+    echo "[ERROR] Services not ready after $max_attempts attempts"
+    return 1
+}
+
+# 函数：显示测试步骤
+show_step() {
+    echo ""
+    echo "🔄 Step $1: $2"
+    echo "----------------------------------------"
+}
+
+# 函数：验证步骤结果
+verify_step() {
+    if [ $? -eq 0 ]; then
+        echo "✅ $1 - SUCCESS"
+    else
+        echo "❌ $1 - FAILED"
+        exit 1
+    fi
+}
+
+# 开始端到端测试
+show_step "1" "Bootstrap - Initialize environment"
+./scripts/01_bootstrap.sh
+verify_step "Bootstrap"
+
+show_step "2" "Startup - Start all services"
+./scripts/02_up.sh
+verify_step "Service startup"
+
+# 等待服务完全就绪
+wait_for_services || exit 1
+
+# 发送告警数据
+show_step "3" "Add alerts - Send test alerts to Alertmanager"
+./scripts/03_alertmanager_add_alert.sh
+verify_step "Send test alerts"
+
+# 查询告警数据
+show_step "4" "Verify data - Query Alertmanager"
+./scripts/04_query_alerts.sh
+verify_step "Data verification"
+
+
+# 检查服务健康状态
+show_step "Health" "Check service health"
+echo "[INFO] Checking service health..."
+
+# 检查 Alertmanager 状态
+if curl -fs "http://localhost:9093/api/v2/status" >/dev/null 2>&1; then
+    am_status="available"
+    echo "✅ Alertmanager status: $am_status"
+else
+    am_status="unavailable"
+    echo "⚠️  Alertmanager status: $am_status"
+fi
+verify_step "Service health check"
+
+# 清理环境
+show_step "5" "Cleanup - Stop all services"
+./scripts/05_down.sh
+verify_step "Service cleanup"
+
+# 计算总测试时间
+test_end_time=$(date +%s)
+total_time=$((test_end_time - test_start_time))
+
+echo ""
+echo "======================================="
+echo "🎉 END-TO-END TEST COMPLETED SUCCESSFULLY!"
+echo "======================================="
+echo "📊 Test Summary:"
+echo "   • Total time: ${total_time}s"
+echo "   • Alertmanager status: $am_status"
+echo "   • All services started and stopped successfully"
+echo ""
+echo "✅ The ARGUS Alert system is working correctly!"
+echo ""

src/bind/build/Dockerfile

@@ -26,6 +26,7 @@ RUN apt-get update && \
     apt-get install -y \
     bind9 \
     bind9utils \
+    dnsutils \
     bind9-doc \
     supervisor \
     net-tools \

src/bind/build/dns-monitor.sh

@@ -17,6 +17,9 @@ log_message() {
 
 log_message "DNS监控脚本启动"
 
+log_message "删除DNS备份文件（如果存在）"
+rm -f $DNS_BACKUP
+
 while true; do
     if [ -f "$DNS_CONF" ]; then
         if [ -f "$DNS_BACKUP" ]; then

src/log/fluent-bit/build/start-fluent-bit.sh

@@ -1,47 +1,96 @@
 #!/bin/bash
 set -euo pipefail
 
-echo "[INFO] Starting Fluent Bit setup in Ubuntu container..."
+echo "[INFO] Starting Fluent Bit setup in Ubuntu container (offline-first)..."
 
-# 安装必要的工具
-echo "[INFO] Installing required packages..."
 export DEBIAN_FRONTEND=noninteractive
-apt-get update -qq
-apt-get install -y -qq curl
 
-# 解压bundle到/tmp
-echo "[INFO] Extracting fluent-bit bundle..."
-cp -r /private/etc /tmp
-cp -r /private/packages /tmp
-cd /tmp
+# Stage bundle to /tmp (read-only mount under /private)
+echo "[INFO] Staging fluent-bit bundle..."
+rm -rf /tmp/flb && mkdir -p /tmp/flb
+cp -r /private/etc /tmp/flb/
+mkdir -p /tmp/flb/packages
+cp -r /private/packages/* /tmp/flb/packages/ 2>/dev/null || true
 
-# 安装 Fluent Bit 从 deb 包
-echo "[INFO] Installing Fluent Bit from deb package..."
-dpkg -i /tmp/packages/fluent-bit_3.1.9_amd64.deb || true
-apt-get install -f -y -qq  # 解决依赖问题
+# Helper: check and install a local deb if not already satisfied
+ensure_lib() {
+  local soname="$1"; shift
+  local pattern="$1"; shift
+  if ldconfig -p 2>/dev/null | grep -q "$soname"; then
+    echo "[OK] $soname already present"
+    return 0
+  fi
+  local deb="$(ls /tmp/flb/packages/$pattern 2>/dev/null | head -n1 || true)"
+  if [[ -n "$deb" ]]; then
+    echo "[INFO] Installing local dependency: $(basename "$deb")"
+    dpkg -i "$deb" >/dev/null 2>&1 || true
+  else
+    echo "[WARN] Local deb for $soname not found (pattern=$pattern)"
+  fi
+  if ! ldconfig -p 2>/dev/null | grep -q "$soname"; then
+    echo "[WARN] $soname still missing after local install; attempting apt fallback"
+    apt-get update -qq || true
+    case "$soname" in
+      libpq.so.5) apt-get install -y -qq libpq5 || true ;;
+      libyaml-0.so.2) apt-get install -y -qq libyaml-0-2 || true ;;
+    esac
+  fi
+  ldconfig 2>/dev/null || true
+}
+
+# Offline-first: satisfy runtime deps from local debs, fallback to apt only if necessary
+ensure_lib "libpq.so.5" "libpq5_*_amd64.deb"
+ensure_lib "libyaml-0.so.2" "libyaml-0-2_*_amd64.deb"
+ensure_lib "libsasl2.so.2" "libsasl2-2_*_amd64.deb"
+ensure_lib "libldap-2.5.so.0" "libldap-2.5-0_*_amd64.deb"
+
+# Install fluent-bit main package from local bundle
+FLB_DEB="$(ls /tmp/flb/packages/fluent-bit_*_amd64.deb 2>/dev/null | head -n1 || true)"
+if [[ -z "$FLB_DEB" ]]; then
+  echo "[ERROR] fluent-bit deb not found under /private/packages" >&2
+  exit 1
+fi
+echo "[INFO] Installing Fluent Bit: $(basename "$FLB_DEB")"
+dpkg -i "$FLB_DEB" >/dev/null 2>&1 || true
+
+# If dpkg reported unresolved dependencies, try apt -f only as last resort
+if ! command -v /opt/fluent-bit/bin/fluent-bit >/dev/null 2>&1; then
+  echo "[WARN] fluent-bit binary missing after dpkg; attempting apt --fix-broken"
+  apt-get install -f -y -qq || true
+fi
+
+# Ensure runtime library dependencies are satisfied (libsasl2, libldap are required via libpq/curl)
+MISSING=$(ldd /opt/fluent-bit/bin/fluent-bit 2>/dev/null | awk '/not found/{print $1}' | xargs -r echo || true)
+if [[ -n "$MISSING" ]]; then
+  echo "[WARN] missing shared libs: $MISSING"
+  apt-get update -qq || true
+  apt-get install -y -qq libsasl2-2 libldap-2.5-0 || true
+  apt-get install -f -y -qq || true
+fi
 
-# 验证 Fluent Bit 可以运行
 echo "[INFO] Fluent Bit version:"
-/opt/fluent-bit/bin/fluent-bit --version
+/opt/fluent-bit/bin/fluent-bit --version || { echo "[ERROR] fluent-bit not installed or libraries missing" >&2; exit 1; }
 
-# 创建配置目录
+# Place configuration
 mkdir -p /etc/fluent-bit
-cp -r /tmp/etc/* /etc/fluent-bit/
+cp -r /tmp/flb/etc/* /etc/fluent-bit/
 
-# 创建日志和缓冲区目录
+# Create logs/buffers dirs
 mkdir -p /logs/train /logs/infer /buffers
 chmod 755 /logs/train /logs/infer /buffers
 
-# 等待 Elasticsearch 就绪
-echo "[INFO] Waiting for Elasticsearch to be ready..."
-while ! curl -fs http://${ES_HOST}:${ES_PORT}/_cluster/health >/dev/null 2>&1; do
-    echo "  Waiting for ES at ${ES_HOST}:${ES_PORT}..."
-    sleep 5
+# Wait for Elasticsearch via bash /dev/tcp to avoid curl dependency
+echo "[INFO] Waiting for Elasticsearch to be ready (tcp ${ES_HOST}:${ES_PORT})..."
+for i in $(seq 1 120); do
+  if exec 3<>/dev/tcp/${ES_HOST}/${ES_PORT}; then
+    exec 3<&- 3>&-
+    echo "[INFO] Elasticsearch is ready"
+    break
+  fi
+  [[ $i -eq 120 ]] && { echo "[ERROR] ES not reachable" >&2; exit 1; }
+  sleep 1
 done
-echo "[INFO] Elasticsearch is ready"
 
-# 启动 Fluent Bit
 echo "[INFO] Starting Fluent Bit with configuration from /etc/fluent-bit/"
 echo "[INFO] Command: /opt/fluent-bit/bin/fluent-bit --config=/etc/fluent-bit/fluent-bit.conf"
-exec /opt/fluent-bit/bin/fluent-bit \
-    --config=/etc/fluent-bit/fluent-bit.conf
+exec /opt/fluent-bit/bin/fluent-bit --config=/etc/fluent-bit/fluent-bit.conf

src/log/kibana/build/kibana-post-start.sh

@@ -2,7 +2,7 @@
 set -euo pipefail
 
 ES_HOST="${ELASTICSEARCH_HOSTS:-http://es:9200}"
-KB_HOST="http://localhost:5601"
+KB_HOST="${KB_HOST:-http://127.0.0.1:5601}"
 
 echo "[INFO] Starting Kibana post-start configuration..."
 
@@ -83,50 +83,37 @@ fix_replicas_idempotent() {
 }
 
 # 幂等创建数据视图
+create_or_ensure_data_view() {
+    local name="$1"
+    local title="$2"
+
+    local list_response
+    list_response=$(curl -fsS "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null || echo "")
+
+    if [ -z "$list_response" ]; then
+        echo "[WARN] Failed to list data views, skipping creation check for $title"
+        return
+    fi
+
+    if echo "$list_response" | grep -Fq "\"title\":\"$title\""; then
+        echo "[INFO] Data view $title already exists, skipping"
+        return
+    fi
+
+    echo "[INFO] Creating data view for $title indices (allowNoIndex)"
+
+    curl -fsS -X POST "$KB_HOST/api/data_views/data_view?allowNoIndex=true" \
+        -H 'kbn-xsrf: true' \
+        -H 'Content-Type: application/json' \
+        -d "{\"data_view\":{\"name\":\"$name\",\"title\":\"$title\",\"timeFieldName\":\"@timestamp\",\"allowNoIndex\":true}}" \
+        >/dev/null && echo "[OK] Created $name data view" || echo "[WARN] Failed to create $name data view"
+}
+
 create_data_views_idempotent() {
     echo "[INFO] Checking and creating data views..."
 
-    # 检查是否存在匹配的索引
-    local train_indices=$(curl -s "$ES_HOST/_cat/indices/train-*?h=index" 2>/dev/null | wc -l || echo "0")
-    local infer_indices=$(curl -s "$ES_HOST/_cat/indices/infer-*?h=index" 2>/dev/null | wc -l || echo "0")
-
-    # 创建 train 数据视图
-    if [ "$train_indices" -gt 0 ]; then
-        # 检查数据视图是否已存在
-        local train_exists=$(curl -s "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null | grep '"title":"train-\*"' | wc -l )
-
-        if [ "$train_exists" -eq 0 ]; then
-            echo "[INFO] Creating data view for train-* indices"
-            curl -fsS -X POST "$KB_HOST/api/data_views/data_view" \
-                -H 'kbn-xsrf: true' \
-                -H 'Content-Type: application/json' \
-                -d '{"data_view":{"name":"train","title":"train-*","timeFieldName":"@timestamp"}}' \
-                >/dev/null && echo "[OK] Created train data view" || echo "[WARN] Failed to create train data view"
-        else
-            echo "[INFO] Train data view already exists, skipping"
-        fi
-    else
-        echo "[INFO] No train-* indices found, skipping train data view creation"
-    fi
-
-    # 创建 infer 数据视图
-    if [ "$infer_indices" -gt 0 ]; then
-        # 检查数据视图是否已存在
-        local infer_exists=$(curl -s "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null | grep '"title":"infer-\*"' | wc -l )
-
-        if [ "$infer_exists" -eq 0 ]; then
-            echo "[INFO] Creating data view for infer-* indices"
-            curl -fsS -X POST "$KB_HOST/api/data_views/data_view" \
-                -H 'kbn-xsrf: true' \
-                -H 'Content-Type: application/json' \
-                -d '{"data_view":{"name":"infer","title":"infer-*","timeFieldName":"@timestamp"}}' \
-                >/dev/null && echo "[OK] Created infer data view" || echo "[WARN] Failed to create infer data view"
-        else
-            echo "[INFO] Infer data view already exists, skipping"
-        fi
-    else
-        echo "[INFO] No infer-* indices found, skipping infer data view creation"
-    fi
+    create_or_ensure_data_view "train" "train-*"
+    create_or_ensure_data_view "infer" "infer-*"
 }
 
 # 主逻辑

src/log/tests/scripts/01_bootstrap.sh

@@ -32,3 +32,42 @@ fi
 
 echo "[OK] 初始化完成: private/argus/log/{elasticsearch,kibana}"
 echo "[INFO] Fluent-bit files should be in fluent-bit/ directory"
+
+# 准备 Fluent Bit 离线依赖（从 metric all-in-one-full 复制 deb 到 ../fluent-bit/build/packages）
+FLB_BUILD_PACKAGES_DIR="$root/../fluent-bit/build/packages"
+mkdir -p "$FLB_BUILD_PACKAGES_DIR"
+for deb in \
+  "$project_root/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libyaml-0-2_"*_amd64.deb \
+  "$project_root/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libpq5_"*_amd64.deb ; do
+  if ls $deb >/dev/null 2>&1; then
+    for f in $deb; do
+      base="$(basename "$f")"
+      if [[ ! -f "$FLB_BUILD_PACKAGES_DIR/$base" ]]; then
+        cp "$f" "$FLB_BUILD_PACKAGES_DIR/"
+        echo "  [+] copied $base"
+      fi
+    done
+  fi
+done
+
+# 额外：从 all-in-one-full 的 ubuntu22/curl.tar.gz 解包必要依赖（libsasl2/ldap），便于离线安装
+CURLOPT_TAR="$project_root/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/curl.tar.gz"
+if [[ -f "$CURLOPT_TAR" ]]; then
+  tmpdir=$(mktemp -d)
+  if tar -xzf "$CURLOPT_TAR" -C "$tmpdir" 2>/dev/null; then
+    for p in \
+      libsasl2-2_*_amd64.deb \
+      libsasl2-modules-db_*_amd64.deb \
+      libldap-2.5-0_*_amd64.deb \
+      libidn2-0_*_amd64.deb \
+      libbrotli1_*_amd64.deb \
+      libssl3_*_amd64.deb ; do
+      src=$(ls "$tmpdir"/curl/$p 2>/dev/null | head -n1 || true)
+      if [[ -n "$src" ]]; then
+        base="$(basename "$src")"
+        [[ -f "$FLB_BUILD_PACKAGES_DIR/$base" ]] || cp "$src" "$FLB_BUILD_PACKAGES_DIR/" && echo "  [+] staged $base"
+      fi
+    done
+  fi
+  rm -rf "$tmpdir"
+fi

src/log/tests/scripts/e2e_test.sh

@@ -115,20 +115,32 @@ show_step "Health" "Check service health"
 echo "[INFO] Checking service health..."
 
 # 检查 Elasticsearch 健康状态
+health_check_ok=1
 es_health=$(curl -s "http://localhost:9200/_cluster/health" | grep -o '"status":"[^"]*"' | cut -d'"' -f4)
 if [ "$es_health" = "green" ] || [ "$es_health" = "yellow" ]; then
     echo "✅ Elasticsearch health: $es_health"
 else
     echo "❌ Elasticsearch health: $es_health"
+    health_check_ok=0
 fi
 
 # 检查 Kibana 状态
 if curl -fs "http://localhost:5601/api/status" >/dev/null 2>&1; then
     kb_status="available"
     echo "✅ Kibana status: $kb_status"
+
+    data_views_json=$(curl -fs "http://localhost:5601/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null || true)
+    if echo "$data_views_json" | grep -F '"title":"train-*"' >/dev/null 2>&1 && \
+       echo "$data_views_json" | grep -F '"title":"infer-*"' >/dev/null 2>&1; then
+        echo "✅ Kibana data views: train-* and infer-* present"
+    else
+        echo "❌ Kibana data views missing: train-* or infer-*"
+        health_check_ok=0
+    fi
 else
     kb_status="unavailable"
     echo "⚠️  Kibana status: $kb_status"
+    health_check_ok=0
 fi
 
 # 检查 Fluent-Bit 指标
@@ -139,6 +151,13 @@ if [ "$fb_host01_uptime" -gt 0 ] && [ "$fb_host02_uptime" -gt 0 ]; then
     echo "✅ Fluent-Bit services: host01 uptime=${fb_host01_uptime}s, host02 uptime=${fb_host02_uptime}s"
 else
     echo "⚠️  Fluent-Bit services: host01 uptime=${fb_host01_uptime}s, host02 uptime=${fb_host02_uptime}s"
+    health_check_ok=0
+fi
+
+if [ "$health_check_ok" -eq 1 ]; then
+    true
+else
+    false
 fi
 
 verify_step "Service health check"

src/master/scripts/build_images.sh

@@ -3,12 +3,13 @@ set -euo pipefail
 
 usage() {
   cat >&2 <<'USAGE'
-Usage: $0 [--intranet] [--offline] [--tag <image_tag>]
+Usage: $0 [--intranet] [--offline] [--tag <image_tag>] [--no-cache]
 
 Options:
   --intranet           使用指定的 PyPI 镜像源（默认清华镜像）。
   --offline            完全离线构建，依赖 offline_wheels/ 目录中的离线依赖包。
   --tag <image_tag>    自定义镜像标签，默认 argus-master:latest。
+  --no-cache           不使用 Docker 构建缓存。
 USAGE
 }
 
@@ -19,6 +20,7 @@ IMAGE_TAG="${IMAGE_TAG:-argus-master:latest}"
 DOCKERFILE="src/master/Dockerfile"
 BUILD_ARGS=()
 OFFLINE_MODE=0
+NO_CACHE=0
 
 source "$PROJECT_ROOT/scripts/common/build_user.sh"
 load_build_user
@@ -45,6 +47,11 @@ while [[ "$#" -gt 0 ]]; do
       IMAGE_TAG="$2"
       shift 2
       ;;
+    --no-cache)
+      NO_CACHE=1
+      BUILD_ARGS+=("--no-cache")
+      shift
+      ;;
     -h|--help)
       usage
       exit 0

src/metric/.gitignore

@@ -4,4 +4,4 @@
 /client-plugins/demo-all-in-one/publish/
 /client-plugins/demo-all-in-one/checklist
 /client-plugins/demo-all-in-one/VERSION
-/client-plugins/all-in-one-full/
+/client-plugins/all-in-one-full/artifact/

src/metric/client-plugins/all-in-one-demo/scripts/check_health.sh

@@ -4,6 +4,15 @@
 
 set -e
 
+# PID 文件检测，防止重复执行
+PIDFILE="/var/run/check_health.pid"
+if [ -f "$PIDFILE" ] && kill -0 $(cat "$PIDFILE") 2>/dev/null; then
+    echo "健康检查脚本已在运行中，跳过本次执行" >&2
+    exit 0
+fi
+echo $$ > "$PIDFILE"
+trap "rm -f $PIDFILE" EXIT
+
 # 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 HEALTH_LOG_FILE="$SCRIPT_DIR/.health_log"

src/metric/client-plugins/all-in-one-demo/scripts/install_artifact.sh

@@ -200,22 +200,22 @@ parse_version_info() {
         VERSION=$(grep '"version"' "$VERSION_FILE_PATH" | sed 's/.*"version": *"\([^"]*\)".*/\1/')
         BUILD_TIME=$(grep '"build_time"' "$VERSION_FILE_PATH" | sed 's/.*"build_time": *"\([^"]*\)".*/\1/')
         
-        # 解析 artifact_list
-        grep -A 100 '"artifact_list"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
+        # 解析 artifact_list（跳过字段名本身）
+        grep -A 100 '"artifact_list"' "$VERSION_FILE_PATH" | grep -v '"artifact_list"' | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)":\s*"[^"]*".*/\1/')
             version=$(echo "$line" | sed 's/.*"[^"]*":\s*"\([^"]*\)".*/\1/')
             echo "$component:$version" >> "$TEMP_DIR/components.txt"
         done
         
-        # 解析 checksums
-        grep -A 100 '"checksums"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
+        # 解析 checksums（跳过字段名本身）
+        grep -A 100 '"checksums"' "$VERSION_FILE_PATH" | grep -v '"checksums"' | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)":\s*"[^"]*".*/\1/')
             checksum=$(echo "$line" | sed 's/.*"[^"]*":\s*"\([^"]*\)".*/\1/')
             echo "$component:$checksum" >> "$TEMP_DIR/checksums.txt"
         done
         
-        # 解析 install_order
-        grep -A 100 '"install_order"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+"' | while read line; do
+        # 解析 install_order（跳过字段名本身，只取数组元素）
+        grep -A 100 '"install_order"' "$VERSION_FILE_PATH" | grep -v '"install_order"' | grep -E '^\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)".*/\1/')
             echo "$component" >> "$TEMP_DIR/install_order.txt"
         done
@@ -317,6 +317,30 @@ create_install_dirs() {
     log_success "安装目录创建完成: $INSTALL_DIR"
 }
 
+# 获取系统版本
+get_system_version() {
+    if [[ ! -f /etc/os-release ]]; then
+        log_error "无法检测操作系统版本"
+        return 1
+    fi
+    
+    source /etc/os-release
+    
+    # 提取主版本号
+    case "$VERSION_ID" in
+        "20.04")
+            echo "ubuntu20"
+            ;;
+        "22.04")
+            echo "ubuntu22"
+            ;;
+        *)
+            log_warning "未识别的Ubuntu版本: $VERSION_ID，尝试使用ubuntu22"
+            echo "ubuntu22"
+            ;;
+    esac
+}
+
 # 安装系统依赖包
 install_system_deps() {
     log_info "检查系统依赖包..."
@@ -330,21 +354,43 @@ install_system_deps() {
         return 0
     fi
 
-    # 检查是否有tar.gz文件
-    local deps_count=$(find "$deps_dir" -name "*.tar.gz" | wc -l)
-    if [[ $deps_count -eq 0 ]]; then
-        log_info "deps 目录中没有 tar.gz 文件，跳过系统依赖包安装"
-        return 0
+    # 获取系统版本对应的依赖目录
+    local system_version=$(get_system_version)
+    local version_deps_dir="$deps_dir/$system_version"
+    
+    log_info "检测到系统版本: $system_version"
+    
+    # 检查版本特定的依赖目录是否存在
+    if [[ ! -d "$version_deps_dir" ]]; then
+        log_warning "未找到 $system_version 版本的依赖目录: $version_deps_dir"
+        # 回退到旧的逻辑，检查根deps目录
+        local deps_count=$(find "$deps_dir" -name "*.tar.gz" | wc -l)
+        if [[ $deps_count -eq 0 ]]; then
+            log_info "deps 目录中没有 tar.gz 文件，跳过系统依赖包安装"
+            return 0
+        fi
+        version_deps_dir="$deps_dir"
+    else
+        # 检查版本目录中是否有tar.gz文件
+        local deps_count=$(find "$version_deps_dir" -name "*.tar.gz" | wc -l)
+        if [[ $deps_count -eq 0 ]]; then
+            log_info "$system_version 版本目录中没有 tar.gz 文件，跳过系统依赖包安装"
+            return 0
+        fi
     fi
 
-    log_info "找到 $deps_count 个系统依赖包，开始安装..."
+    log_info "找到 $system_version 版本的依赖包，开始安装..."
 
     # 创建临时目录用于解压依赖包
-    local deps_temp_dir="$TEMP_DIR/deps"
+    local deps_temp_dir="${TEMP_DIR:-/tmp}/deps"
     mkdir -p "$deps_temp_dir"
 
+    # 定义要检查的核心依赖
+    local CORE_DEPS=(jq cron curl)
+    local FAILED_DEPS=()
+
     # 处理每个tar.gz文件
-    find "$deps_dir" -name "*.tar.gz" | while read tar_file; do
+    find "$version_deps_dir" -name "*.tar.gz" | while read tar_file; do
         local tar_basename=$(basename "$tar_file")
         local extract_name="${tar_basename%.tar.gz}"
 
@@ -362,40 +408,61 @@ install_system_deps() {
         fi
 
         # 进入解压目录，查找deb包
-        cd "$extract_dir"
-        local deb_count=$(find . -name "*.deb" | wc -l)
+        cd "$extract_dir" || continue
+        local deb_files=(*.deb)
+        if [[ ${#deb_files[@]} -gt 0 ]]; then
+            log_info "  找到 ${#deb_files[@]} 个 deb 包，开始安装..."
 
-        if [[ $deb_count -gt 0 ]]; then
-            log_info "  找到 $deb_count 个 deb 包，开始安装..."
+            for deb in "${deb_files[@]}"; do
+                local pkg_name
+                pkg_name=$(dpkg-deb -f "$deb" Package 2>/dev/null)
 
-            # 1. 先尝试安装所有deb包
-            log_info "  第1步：批量安装deb包..."
-            if dpkg -i *.deb 2>/dev/null; then
-                log_success "  所有deb包安装成功"
-            else
-                log_warning "  部分deb包安装失败，可能存在依赖问题"
-                
-                # 2. 使用apt-get修复依赖
-                log_info "  第2步：修复依赖关系..."
-                if apt-get install -f -y; then
-                    log_success "  依赖关系修复完成"
-                else
-                    log_error "  依赖关系修复失败"
-                    # 继续处理其他包，不退出
+                # 如果已安装，则跳过
+                if dpkg -s "$pkg_name" &>/dev/null; then
+                    log_success "  $pkg_name 已安装，跳过"
+                    continue
                 fi
-            fi
+
+                # 尝试安装
+                log_info "  安装 $pkg_name..."
+                if DEBIAN_FRONTEND=noninteractive dpkg -i "$deb" &>/dev/null; then
+                    log_success "  $pkg_name 安装成功"
+                else
+                    log_warning "  $pkg_name 安装失败，尝试修复依赖..."
+                    if DEBIAN_FRONTEND=noninteractive apt-get install -f -y &>/dev/null; then
+                        if dpkg -s "$pkg_name" &>/dev/null; then
+                            log_success "  $pkg_name 修复安装成功"
+                        else
+                            log_error "  $pkg_name 仍未安装成功"
+                            FAILED_DEPS+=("$pkg_name")
+                        fi
+                    else
+                        log_error "  $pkg_name 自动修复失败"
+                        FAILED_DEPS+=("$pkg_name")
+                    fi
+                fi
+            done
         else
             log_info "  $tar_basename 中没有找到deb包，跳过"
         fi
 
         # 返回到依赖临时目录
-        cd "$deps_temp_dir"
+        cd "$deps_temp_dir" || continue
     done
 
     # 检查并启动 cron 服务
     start_cron_service
 
-    log_success "系统依赖包安装完成"
+    # 总结安装结果
+    if [[ ${#FAILED_DEPS[@]} -gt 0 ]]; then
+        log_error "以下系统依赖未能成功安装，安装终止，请手动安装后重试："
+        for f in "${FAILED_DEPS[@]}"; do
+            echo "  - $f"
+        done
+        exit 1
+    else
+        log_success "系统依赖包安装完成，全部就绪"
+    fi
 }
 
 # 启动 cron 服务
@@ -637,6 +704,18 @@ EOF
     log_success "安装记录已创建: $install_record_file"
 }
 
+# 检查cron任务是否已存在
+check_cron_task_exists() {
+    local task_pattern="$1"
+    local temp_cron="$2"
+    
+    if grep -q "$task_pattern" "$temp_cron"; then
+        return 0  # 任务已存在
+    else
+        return 1  # 任务不存在
+    fi
+}
+
 # 设置健康检查定时任务
 setup_health_check_cron() {
     log_info "设置健康检查定时任务..."
@@ -661,7 +740,7 @@ setup_health_check_cron() {
     crontab -l 2>/dev/null > "$temp_cron" || touch "$temp_cron"
     
     # 检查并删除旧的健康检查任务
-    if grep -q "check_health.sh" "$temp_cron"; then
+    if check_cron_task_exists "check_health.sh" "$temp_cron"; then
         log_info "发现旧的健康检查定时任务，正在更新..."
         # 删除所有包含check_health.sh的行
         grep -v "check_health.sh" "$temp_cron" > "$temp_cron.new"
@@ -716,7 +795,7 @@ setup_dns_sync_cron() {
     crontab -l 2>/dev/null > "$temp_cron" || touch "$temp_cron"
     
     # 检查并删除旧的 DNS 同步任务
-    if grep -q "sync_dns.sh" "$temp_cron"; then
+    if check_cron_task_exists "sync_dns.sh" "$temp_cron"; then
         log_info "发现旧的 DNS 同步定时任务，正在更新..."
         # 删除所有包含sync_dns.sh的行
         grep -v "sync_dns.sh" "$temp_cron" > "$temp_cron.new"
@@ -724,16 +803,15 @@ setup_dns_sync_cron() {
         log_info "旧的 DNS 同步定时任务已删除"
     fi
     
-    # 添加新的定时任务（每30秒执行一次）
+    # 添加新的定时任务（每1分钟执行一次）
     # 直接使用版本目录中的 DNS 同步脚本
     echo "# Argus-Metrics DNS 同步定时任务" >> "$temp_cron"
     echo "* * * * * $sync_dns_script >> $INSTALL_DIR/.dns_sync.log 2>&1" >> "$temp_cron"
-    echo "* * * * * sleep 30; $sync_dns_script >> $INSTALL_DIR/.dns_sync.log 2>&1" >> "$temp_cron"
     
     # 安装新的crontab
     if crontab "$temp_cron"; then
         log_success "DNS 同步定时任务设置成功"
-        log_info "  执行频率: 每30秒"
+        log_info "  执行频率: 每1分钟"
         log_info "  日志文件: $INSTALL_DIR/.dns_sync.log"
         log_info "  查看定时任务: crontab -l"
         log_info "  删除定时任务: crontab -e"
@@ -771,7 +849,7 @@ setup_version_check_cron() {
     crontab -l > "$temp_cron" 2>/dev/null || touch "$temp_cron"
     
     # 检查是否已存在版本校验定时任务
-    if grep -q "check_version.sh" "$temp_cron"; then
+    if check_cron_task_exists "check_version.sh" "$temp_cron"; then
         log_info "发现旧的版本校验定时任务，正在更新..."
         # 删除所有包含check_version.sh的行
         grep -v "check_version.sh" "$temp_cron" > "$temp_cron.new"
@@ -824,7 +902,7 @@ setup_restart_cron() {
     crontab -l > "$temp_cron" 2>/dev/null || touch "$temp_cron"
     
     # 检查是否已存在自动重启定时任务
-    if grep -q "restart_unhealthy.sh" "$temp_cron"; then
+    if check_cron_task_exists "restart_unhealthy.sh" "$temp_cron"; then
         log_info "发现旧的自动重启定时任务，正在更新..."
         # 删除所有包含restart_unhealthy.sh的行
         grep -v "restart_unhealthy.sh" "$temp_cron" > "$temp_cron.new"
@@ -885,9 +963,9 @@ main() {
     check_system
     find_version_file
     create_install_dirs
+    install_system_deps     
     parse_version_info    
     verify_checksums
-    install_system_deps
     install_components
     copy_config_files
     create_install_record
@@ -895,6 +973,20 @@ main() {
     setup_dns_sync_cron
     setup_version_check_cron
     setup_restart_cron
+    
+    # 注释掉立即执行健康检查，避免与cron任务重复执行
+    # log_info "立即执行一次健康检查..."
+    # local check_health_script="$INSTALL_DIR/check_health.sh"
+    # if [[ -f "$check_health_script" ]]; then
+    #     if "$check_health_script" >> "$INSTALL_DIR/.health_check.log" 2>&1; then
+    #         log_success "健康检查执行完成"
+    #     else
+    #         log_warning "健康检查执行失败，请检查日志: $INSTALL_DIR/.health_check.log"
+    #     fi
+    # else
+    #     log_warning "健康检查脚本不存在: $check_health_script"
+    # fi
+    
     show_install_info
 }
 

src/metric/client-plugins/all-in-one-demo/scripts/publish_artifact.sh

@@ -29,26 +29,68 @@ log_error() {
 show_help() {
     echo "Argus-Metric Artifact 发布脚本"
     echo
-    echo "用法: $0 <版本号>"
+    echo "用法: $0 <版本号> [选项]"
     echo
     echo "参数:"
-    echo "  <版本号>    要发布的版本号，对应 artifact 目录中的版本"
+    echo "  <版本号>              要发布的版本号，对应 artifact 目录中的版本"
+    echo
+    echo "选项:"
+    echo "  --output-dir <路径>   指定输出目录 (默认: /private/argus/ftp/share/)"
+    echo "  --owner <uid:gid>     指定文件所有者 (默认: 2133:2015)"
+    echo "  -h, --help           显示此帮助信息"
     echo
     echo "示例:"
-    echo "  $0 1.20.0    # 发布 1.20.0 版本"
+    echo "  $0 1.20.0                                    # 使用默认配置发布"
+    echo "  $0 1.20.0 --output-dir /tmp/publish          # 指定输出目录"
+    echo "  $0 1.20.0 --owner 1000:1000                  # 指定文件所有者"
+    echo "  $0 1.20.0 --output-dir /srv/ftp --owner root:root  # 同时指定两者"
     echo
 }
 
-# 检查参数
-if [[ $# -ne 1 ]]; then
+# 默认配置
+DEFAULT_PUBLISH_DIR="/private/argus/ftp/share/"
+DEFAULT_OWNER="2133:2015"
+
+# 解析参数
+VERSION=""
+PUBLISH_DIR="$DEFAULT_PUBLISH_DIR"
+OWNER="$DEFAULT_OWNER"
+
+while [[ $# -gt 0 ]]; do
+    case $1 in
+        -h|--help)
+            show_help
+            exit 0
+            ;;
+        --output-dir)
+            PUBLISH_DIR="$2"
+            shift 2
+            ;;
+        --owner)
+            OWNER="$2"
+            shift 2
+            ;;
+        *)
+            if [[ -z "$VERSION" ]]; then
+                VERSION="$1"
+                shift
+            else
+                log_error "未知参数: $1"
+                show_help
+                exit 1
+            fi
+            ;;
+    esac
+done
+
+# 检查版本号是否提供
+if [[ -z "$VERSION" ]]; then
     log_error "请提供版本号参数"
     show_help
     exit 1
 fi
 
-VERSION="$1"
 ARTIFACT_DIR="artifact/$VERSION"
-PUBLISH_DIR="/Users/sundapeng/Project/nlp/aiops/client-plugins/all-in-one/publish/"
 
 # 检查版本目录是否存在
 if [[ ! -d "$ARTIFACT_DIR" ]]; then
@@ -57,11 +99,32 @@ if [[ ! -d "$ARTIFACT_DIR" ]]; then
 fi
 
 log_info "开始发布版本: $VERSION"
+log_info "输出目录: $PUBLISH_DIR"
+log_info "文件所有者: $OWNER"
 
 # 确保发布目录存在
 log_info "确保发布目录存在: $PUBLISH_DIR"
 mkdir -p "$PUBLISH_DIR"
 
+IFS=':' read -r OWNER_UID OWNER_GID <<< "$OWNER"
+if [[ -z "$OWNER_UID" || -z "$OWNER_GID" ]]; then
+    log_error "--owner 格式不正确，应为 uid:gid"
+    exit 1
+fi
+
+CURRENT_UID=$(id -u)
+CURRENT_GID=$(id -g)
+if [[ "$OWNER_UID" != "$CURRENT_UID" || "$OWNER_GID" != "$CURRENT_GID" ]]; then
+    if [[ "$CURRENT_UID" -ne 0 ]]; then
+        log_error "当前用户 (${CURRENT_UID}:${CURRENT_GID}) 无法设置所有者为 ${OWNER_UID}:${OWNER_GID}"
+        log_error "请以目标用户运行脚本或预先调整目录权限"
+        exit 1
+    fi
+    NEED_CHOWN=true
+else
+    NEED_CHOWN=false
+fi
+
 # 创建临时目录用于打包
 TEMP_PACKAGE_DIR="/tmp/argus-metric-package-$$"
 mkdir -p "$TEMP_PACKAGE_DIR"
@@ -167,17 +230,28 @@ cd "$TEMP_PACKAGE_DIR"
 tar -czf "$PUBLISH_DIR/$TAR_NAME" *
 cd - > /dev/null
 
+if [[ "$NEED_CHOWN" == true ]]; then
+    log_info "设置文件所有者为: $OWNER"
+    chown "$OWNER" "$PUBLISH_DIR/$TAR_NAME"
+fi
+
 # 清理临时目录
 rm -rf "$TEMP_PACKAGE_DIR"
 
 # 更新 LATEST_VERSION 文件
 log_info "更新 LATEST_VERSION 文件..."
 echo "$VERSION" > "$PUBLISH_DIR/LATEST_VERSION"
+if [[ "$NEED_CHOWN" == true ]]; then
+    chown "$OWNER" "$PUBLISH_DIR/LATEST_VERSION"
+fi
 
 # 复制 DNS 配置文件到发布目录根目录（直接从 config 目录复制）
 if [[ -f "config/dns.conf" ]]; then
     log_info "复制 DNS 配置文件到发布目录根目录..."
     cp "config/dns.conf" "$PUBLISH_DIR/"
+    if [[ "$NEED_CHOWN" == true ]]; then
+        chown "$OWNER" "$PUBLISH_DIR/dns.conf"
+    fi
     log_success "DNS 配置文件复制完成: $PUBLISH_DIR/dns.conf"
 else
     log_warning "未找到 config/dns.conf 文件，跳过 DNS 配置文件复制"
@@ -187,6 +261,9 @@ fi
 if [[ -f "scripts/setup.sh" ]]; then
     log_info "复制 setup.sh 到发布目录..."
     cp "scripts/setup.sh" "$PUBLISH_DIR/"
+    if [[ "$NEED_CHOWN" == true ]]; then
+        chown "$OWNER" "$PUBLISH_DIR/setup.sh"
+    fi
 fi
 
 # 显示发布结果

src/metric/client-plugins/all-in-one-demo/scripts/restart_unhealthy.sh

@@ -2,6 +2,15 @@
 
 # 此脚本会检查各组件的健康状态，并重启不健康的组件
 
+# PID 文件检测，防止重复执行
+PIDFILE="/var/run/restart_unhealthy.pid"
+if [ -f "$PIDFILE" ] && kill -0 $(cat "$PIDFILE") 2>/dev/null; then
+    echo "自动重启脚本已在运行中，跳过本次执行" >&2
+    exit 0
+fi
+echo $$ > "$PIDFILE"
+trap "rm -f $PIDFILE" EXIT
+
 # 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 INSTALL_RECORD_FILE="$SCRIPT_DIR/.install_record"

src/metric/client-plugins/all-in-one-demo/scripts/sync_dns.sh

@@ -1,244 +1,143 @@
 #!/bin/bash
-
-# DNS 同步脚本
-# 比较 FTP 根目录的 dns.conf 和本地的 dns.conf，如果有变化则同步到 /etc/resolv.conf
-
 set -e
 
-# 颜色定义
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-BLUE='\033[0;34m'
-NC='\033[0m' # No Color
+# 颜色
+RED='\033[0;31m'; GREEN='\033[0;32m'; YELLOW='\033[1;33m'; BLUE='\033[0;34m'; NC='\033[0m'
 
-# 日志函数 - 输出到 stderr 避免影响函数返回值
-log_info() {
-    echo -e "${BLUE}[INFO]${NC} $1" >&2
-}
+# 日志函数
+log_info() { echo -e "${BLUE}[INFO]${NC} $1" >&2; }
+log_success() { echo -e "${GREEN}[SUCCESS]${NC} $1" >&2; }
+log_warning() { echo -e "${YELLOW}[WARNING]${NC} $1" >&2; }
+log_error() { echo -e "${RED}[ERROR]${NC} $1" >&2; }
 
-log_success() {
-    echo -e "${GREEN}[SUCCESS]${NC} $1" >&2
-}
-
-log_warning() {
-    echo -e "${YELLOW}[WARNING]${NC} $1" >&2
-}
-
-log_error() {
-    echo -e "${RED}[ERROR]${NC} $1" >&2
-}
-
-# 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 LOCAL_DNS_CONF="/opt/argus-metric/dns.conf"
-REMOTE_DNS_CONF_URL=""
 RESOLV_CONF="/etc/resolv.conf"
+ALT_RESOLV_CONF="/run/resolv.conf"
 LOG_FILE="/opt/argus-metric/.dns_sync.log"
+REMOTE_DNS_CONF_URL=""
 
-# 从环境变量或配置文件获取 FTP 服务器信息
+# 获取 FTP 配置
 get_ftp_config() {
-    # 优先从环境变量获取配置
     log_info "获取 FTP 配置信息..."
-    
-    # 如果环境变量中没有设置，则尝试从配置文件读取
     if [[ -z "$FTP_SERVER" || -z "$FTP_USER" || -z "$FTP_PASSWORD" ]]; then
-        local config_file="$SCRIPT_DIR/config.env"
-        if [[ -f "$config_file" ]]; then
-            log_info "从配置文件读取 FTP 配置: $config_file"
-            source "$config_file"
-        fi
-    else
-        log_info "使用环境变量中的 FTP 配置"
+        [[ -f "$SCRIPT_DIR/config.env" ]] && source "$SCRIPT_DIR/config.env"
     fi
-    
-    # 设置默认值（如果环境变量和配置文件都没有设置）
     FTP_SERVER="${FTP_SERVER:-localhost}"
     FTP_USER="${FTP_USER:-ftpuser}"
     FTP_PASSWORD="${FTP_PASSWORD:-ZGClab1234!}"
-    
-    # 构建远程 DNS 配置文件 URL
     REMOTE_DNS_CONF_URL="ftp://${FTP_USER}:${FTP_PASSWORD}@${FTP_SERVER}/dns.conf"
-    
-    log_info "FTP 配置来源: ${FTP_CONFIG_SOURCE:-环境变量/配置文件}"
 }
 
-# 下载远程 DNS 配置文件
+# 下载远程 dns.conf
 download_remote_dns_conf() {
-    local temp_file="/tmp/dns.conf.remote.$$"
-    
-    log_info "从 FTP 服务器下载 DNS 配置文件..."
-    log_info "远程地址: $REMOTE_DNS_CONF_URL"
-    log_info "FTP 服务器: $FTP_SERVER"
-    log_info "FTP 用户: $FTP_USER"
-    
-    # 先测试 FTP 连接
+    local tmp="/tmp/dns.remote.$$"
     log_info "测试 FTP 连接..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/" >/dev/null 2>&1; then
-        log_success "FTP 服务器连接成功"
-    else
-        log_error "无法连接到 FTP 服务器: $FTP_SERVER"
-        log_error "请检查:"
-        log_error "  1. FTP 服务器是否运行"
-        log_error "  2. 网络连接是否正常"
-        log_error "  3. 服务器地址是否正确"
-        return 1
+    if ! curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/" >/dev/null; then
+        log_error "无法连接到 FTP 服务器: $FTP_SERVER"; return 1
     fi
-    
-    # 测试 dns.conf 文件是否存在
-    log_info "检查远程 dns.conf 文件是否存在..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/dns.conf" >/dev/null 2>&1; then
-        log_success "远程 dns.conf 文件存在"
-    else
-        log_error "远程 dns.conf 文件不存在或无法访问"
-        log_error "请检查 FTP 服务器根目录下是否有 dns.conf 文件"
-        return 1
-    fi
-    
-    # 尝试下载文件
-    log_info "开始下载 dns.conf 文件..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sf "ftp://${FTP_SERVER}/dns.conf" -o "$temp_file" 2>/dev/null; then
-        log_success "远程 DNS 配置文件下载成功"
-        echo "$temp_file"
-    else
-        log_error "下载 dns.conf 文件失败"
-        log_error "尝试手动测试命令:"
-        log_error "  curl -u ${FTP_USER}:${FTP_PASSWORD} ftp://${FTP_SERVER}/dns.conf"
-        rm -f "$temp_file"
-        return 1
+    if ! curl -u "${FTP_USER}:${FTP_PASSWORD}" -sf "ftp://${FTP_SERVER}/dns.conf" -o "$tmp" 2>/dev/null; then
+        log_error "下载 dns.conf 失败"; rm -f "$tmp"; return 1
     fi
+    echo "$tmp"
 }
 
-# 比较两个文件是否相同
-compare_files() {
-    local file1="$1"
-    local file2="$2"
+# 文件比较
+compare_files() { diff -q "$1" "$2" >/dev/null 2>&1; }
 
-    if [[ ! -f "$file1" || ! -f "$file2" ]]; then
-        return 1
-    fi
-    
-    # 使用 diff 比较文件内容
-    if diff -q "$file1" "$file2" >/dev/null 2>&1; then
-        return 0  # 文件相同
-    else
-        return 1  # 文件不同
-    fi
+# 从 dns.conf 提取有效 IP
+get_dns_ips() {
+    grep -Eo '^[0-9]{1,3}(\.[0-9]{1,3}){3}$' "$1" | sort -u
 }
 
-# 将 DNS 配置追加到 /etc/resolv.conf
+# 安全更新 resolv.conf（保留符号链接）
 update_resolv_conf() {
-    local dns_conf_file="$1"
+    local dns_conf="$1"
+    local dns_ips
+    mapfile -t dns_ips < <(get_dns_ips "$dns_conf")
+    [[ ${#dns_ips[@]} -eq 0 ]] && { log_warning "未检测到有效 DNS"; return; }
 
-    log_info "更新 /etc/resolv.conf 文件..."
-    
-    # 备份原始文件
-    if [[ -f "$RESOLV_CONF" ]]; then
-        cp "$RESOLV_CONF" "${RESOLV_CONF}.backup.$(date +%Y%m%d_%H%M%S)"
-        log_info "已备份原始 resolv.conf 文件"
+    local target_file="$RESOLV_CONF"
+    if [[ ! -w "$RESOLV_CONF" ]]; then
+        log_warning "/etc/resolv.conf 不可写，使用兜底路径 $ALT_RESOLV_CONF"
+        target_file="$ALT_RESOLV_CONF"
     fi
 
-    # 读取 DNS 配置文件并追加到 resolv.conf
-    while IFS= read -r line; do
-        # 跳过空行和注释行
-        [[ -z "$line" || "$line" =~ ^[[:space:]]*# ]] && continue
+    local temp="/tmp/resolv.new.$$"
+    cp "$target_file" "${target_file}.backup.$(date +%Y%m%d_%H%M%S)" 2>/dev/null || true
+    log_info "更新 DNS 配置文件: $target_file"
 
-        # 验证是否为有效的 IP 地址
-        if [[ "$line" =~ ^[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}$ ]]; then
-            # 检查是否已存在相同的 nameserver 行
-            if ! grep -q "nameserver $line" "$RESOLV_CONF" 2>/dev/null; then
-                echo "nameserver $line" >> "$RESOLV_CONF"
-                log_info "添加 DNS 服务器: $line"
-            else
-                log_info "DNS 服务器已存在，跳过: $line"
-            fi
-        else
-            log_warning "跳过无效的 DNS 地址: $line"
+    # 写入新的 nameserver 行
+    for ip in "${dns_ips[@]}"; do
+        echo "nameserver $ip"
+    done >"$temp"
+
+    # 追加原内容（去掉重复 nameserver）
+    grep -v '^nameserver' "$target_file" >>"$temp" 2>/dev/null || true
+    awk '!a[$0]++' "$temp" >"${temp}.uniq"
+
+    # ⚙️ 使用 cat 原地覆盖，避免 mv 引发 “设备忙”
+    if cat "${temp}.uniq" >"$target_file" 2>/dev/null; then
+        chmod 644 "$target_file"
+        log_success "DNS 更新完成: ${dns_ips[*]}"
+    else
+        log_error "无法写入 $target_file，可能被系统锁定"
+    fi
+
+    rm -f "$temp" "${temp}.uniq"
+}
+
+# 检查 resolv.conf 是否包含 dns.conf 内容
+ensure_dns_in_resolv() {
+    local dns_conf="$1"
+    local dns_ips
+    mapfile -t dns_ips < <(get_dns_ips "$dns_conf")
+    [[ ${#dns_ips[@]} -eq 0 ]] && return
+
+    for ip in "${dns_ips[@]}"; do
+        if ! grep -q "nameserver $ip" "$RESOLV_CONF" 2>/dev/null; then
+            log_warning "检测到 /etc/resolv.conf 缺少 $ip，执行兜底修复"
+            update_resolv_conf "$dns_conf"
+            return
         fi
-    done < "$dns_conf_file"
-    
-    # 设置文件权限
-    chmod 644 "$RESOLV_CONF"
-    
-    log_success "/etc/resolv.conf 文件更新完成"
+    done
+    log_info "/etc/resolv.conf 已包含所有 DNS"
 }
 
-# 记录同步日志
-log_sync() {
-    local message="$1"
-    local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
-    echo "[$timestamp] $message" >> "$LOG_FILE"
-}
+log_sync() { echo "[$(date '+%F %T')] $1" >>"$LOG_FILE"; }
 
-# 主函数
 main() {
     log_info "开始 DNS 同步检查..."
-    log_sync "DNS 同步检查开始"
+    mkdir -p /opt/argus-metric
 
-    # 确保系统目录存在
-    mkdir -p "/opt/argus-metric"
-    
-    # 获取 FTP 配置
     get_ftp_config
+    local remote_file
+    if ! remote_file=$(download_remote_dns_conf); then
+        log_error "下载失败"; log_sync "同步失败"; exit 1
+    fi
 
-    # 检查本地 DNS 配置文件是否存在
     if [[ ! -f "$LOCAL_DNS_CONF" ]]; then
-        log_warning "本地 DNS 配置文件不存在: $LOCAL_DNS_CONF"
-        log_warning "将下载远程配置文件并更新系统 DNS 设置"
-        
-        # 下载远程配置文件
-        if remote_file=$(download_remote_dns_conf); then
-            # 复制到本地
-            cp "$remote_file" "$LOCAL_DNS_CONF"
-            log_success "远程 DNS 配置文件已保存到本地"
-            
-            # 更新 resolv.conf
-            update_resolv_conf "$LOCAL_DNS_CONF"
-            log_sync "首次同步完成，DNS 配置已更新"
-            
-            # 清理临时文件
-            rm -f "$remote_file"
-        else
-            log_error "无法下载远程 DNS 配置文件，同步失败"
-            log_sync "同步失败：无法下载远程配置文件"
-            exit 1
-        fi
+        log_info "本地 dns.conf 不存在，初始化..."
+        cp "$remote_file" "$LOCAL_DNS_CONF"
+        update_resolv_conf "$LOCAL_DNS_CONF"
+        log_sync "首次同步完成"
     else
-        log_info "本地 DNS 配置文件存在: $LOCAL_DNS_CONF"
-        
-        # 下载远程配置文件进行比较
-        if remote_file=$(download_remote_dns_conf); then
-            # 比较文件
-            if compare_files "$LOCAL_DNS_CONF" "$remote_file"; then
-                log_info "DNS 配置文件无变化，无需更新"
-                log_sync "DNS 配置文件无变化"
-            else
-                log_info "检测到 DNS 配置文件有变化，开始同步..."
-                log_sync "检测到 DNS 配置文件变化，开始同步"
-                
-                # 更新本地配置文件
-                cp "$remote_file" "$LOCAL_DNS_CONF"
-                log_success "本地 DNS 配置文件已更新"
-                
-                # 更新 resolv.conf
-                update_resolv_conf "$LOCAL_DNS_CONF"
-                log_sync "DNS 配置同步完成"
-            fi
-            
-            # 清理临时文件
-            rm -f "$remote_file"
+        if compare_files "$LOCAL_DNS_CONF" "$remote_file"; then
+            log_info "dns.conf 无变化"
+            ensure_dns_in_resolv "$LOCAL_DNS_CONF"
+            log_sync "dns.conf 无变化，执行兜底检查"
         else
-            log_error "无法下载远程 DNS 配置文件，跳过本次同步"
-            log_sync "同步失败：无法下载远程配置文件"
-            exit 1
+            log_info "检测到 DNS 配置更新"
+            cp "$remote_file" "$LOCAL_DNS_CONF"
+            update_resolv_conf "$LOCAL_DNS_CONF"
+            log_sync "DNS 配置同步完成"
         fi
     fi
 
-    log_success "DNS 同步检查完成"
-    log_sync "DNS 同步检查完成"
+    rm -f "$remote_file"
+    log_success "DNS 同步流程完成"
 }
 
-# 脚本入口
 if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
     main "$@"
 fi

src/metric/client-plugins/all-in-one-full/README.md

@@ -0,0 +1,59 @@
+# 客户侧组件安装包构建、发布流程
+
+## 第一步：配置版本和组件
+
+首先搞定配置文件：
+
+1. 把 `.checklist.example` 重命名成 `checklist`
+2. 把 `.VERSION.example` 重命名成 `VERSION`
+
+### checklist 文件格式
+```
+# 组件名称 目录路径 版本号 [依赖组件] [安装顺序]
+dcgm-exporter-installer /path/to/dcgm-exporter-installer 1.1.0
+node-exporter-installer /path/to/node-exporter-installer 1.1.0
+```
+
+### VERSION 文件
+设置需要发布的版本号，比如 `1.29.0`
+
+> 建议用 `version-manager.sh` 来管理版本
+
+## 第二步：构建安装包
+
+直接跑脚本：
+```bash
+./package_artifact.sh
+```
+
+构建完的东西会放在 `artifact/` 目录下，按版本分文件夹。
+
+如果版本已经存在了，想要覆盖重新构建：
+```bash
+./package_artifact.sh --force
+```
+
+构建完可以手工测试安装包。
+
+## 第三步：发布安装包
+
+用这个脚本发布：
+```bash
+./publish_artifact.sh
+```
+
+发布后的内容在 `publish/` 目录里，包含：
+- 压缩版本的安装包
+- 一键安装的bash脚本
+
+## 第四步：部署到FTP服务器
+
+把发布的内容上传到FTP服务器，客户端就可以通过一键命令安装：
+
+```bash
+curl -fsSL http://your-ftp-server/install.sh | sh -
+
+curl -fsSL "ftp://ftpuser:{PASSWD}!@10.211.55.4/share/setup.sh" | sudo bash -s -- --server 10.211.55.4 --user ftpuser --password {PASSWD}
+```
+
+这样客户就能直接从FTP服务器下载并安装组件了。

src/metric/client-plugins/all-in-one-full/config/.VERSION.example

@@ -0,0 +1 @@
+1.29.0

src/metric/client-plugins/all-in-one-full/config/.checklist.example

@@ -0,0 +1,3 @@
+# 组件名称 目录路径 版本号 [依赖组件] [安装顺序]
+dcgm-exporter-installer /Users/sundapeng/Project/nlp/aiops/client-plugins/dcgm-exporter-installer 1.1.0
+node-exporter-installer /Users/sundapeng/Project/nlp/aiops/client-plugins/node-exporter-installer 1.1.0

src/metric/client-plugins/all-in-one-full/config/VERSION

@@ -0,0 +1 @@
+1.35.0

src/metric/client-plugins/all-in-one-full/config/checklist

@@ -0,0 +1,5 @@
+# 组件名称 目录路径 版本号 [依赖组件] [安装顺序]
+argus-agent plugins/argus-agent 1.0.0
+node-exporter plugins/node-exporter 1.0.0
+dcgm-exporter plugins/dcgm-exporter 1.0.0
+fluent-bit plugins/fluent-bit 1.0.0

src/metric/client-plugins/all-in-one-full/config/config.env

@@ -0,0 +1,14 @@
+# Elasticsearch
+ES_HOST=es.log.argus.com
+ES_PORT=9200
+
+# Argus-Agent
+# 连接master服务
+MASTER_ENDPOINT=master.argus.com:3000
+# 上报状态间隔描述
+REPORT_INTERVAL_SECONDS=5
+
+# FTP
+FTP_SERVER=172.31.0.40
+FTP_USER=ftpuser
+FTP_PASSWORD=ZGClab1234!

src/metric/client-plugins/all-in-one-full/config/config.env.example

@@ -0,0 +1,8 @@
+# Argus Metric 配置文件示例
+# 复制此文件为 config.env 并根据需要修改配置
+
+# 连接master服务
+MASTER_ENDPOINT=master.argus.com:3000
+
+# 上报状态间隔描述（秒）
+REPORT_INTERVAL_SECONDS=60

src/metric/client-plugins/all-in-one-full/config/dns.conf

@@ -0,0 +1 @@
+172.31.0.2

src/metric/client-plugins/all-in-one-full/plugins/argus-agent/README.md

@@ -0,0 +1,94 @@
+# Argus Agent 插件
+
+这是 Argus Agent 的安装和管理插件，提供了完整的安装、卸载、健康检查功能。
+
+## 文件结构
+
+```
+argus-agent/
+├── bin/
+│   └── argus-agent          # Argus Agent 二进制文件
+├── config/                  # 配置文件目录
+├── install.sh              # 安装脚本
+├── uninstall.sh            # 卸载脚本
+├── check_health.sh         # 健康检查脚本
+├── package.sh              # 打包脚本
+└── README.md               # 说明文档
+```
+
+## 使用方法
+
+### 安装
+
+```bash
+sudo ./install.sh
+```
+
+安装脚本会：
+- 检查系统要求
+- 停止可能运行的服务
+- 安装二进制文件到 `/usr/local/bin/argus-agent`
+- 创建 `argus-agent` 用户
+- 创建配置和数据目录
+- 启动服务并记录 PID
+
+### 卸载
+
+```bash
+sudo ./uninstall.sh
+```
+
+卸载脚本会：
+- 停止所有 argus-agent 进程
+- 删除二进制文件
+- 删除配置和数据目录
+- 清理日志文件
+- 更新安装记录
+
+### 健康检查
+
+```bash
+./check_health.sh
+```
+
+健康检查脚本会：
+- 检查安装记录中的 PID
+- 验证进程是否正在运行
+- 输出 JSON 格式的健康状态
+
+### 打包
+
+```bash
+./package.sh
+```
+
+打包脚本会：
+- 检查所有必要文件
+- 创建时间戳命名的压缩包
+- 输出安装包信息
+
+## 安装后的文件位置
+
+- 二进制文件: `/usr/local/bin/argus-agent`
+- 配置目录: `/etc/argus-agent/`
+- 数据目录: `/var/lib/argus-agent/`
+- 日志文件: `/var/log/argus-agent.log`
+- PID 文件: `/var/run/argus-agent.pid`
+- 安装记录: `/opt/argus-metric/current/.install_record`
+
+## 健康检查输出格式
+
+```json
+{
+  "name": "argus-agent",
+  "status": "health|unhealth",
+  "reason": "状态说明"
+}
+```
+
+## 注意事项
+
+1. 安装和卸载脚本需要 root 权限
+2. 健康检查脚本使用安装记录中的 PID 来验证进程状态
+3. 如果 jq 命令不可用，健康检查会使用简单的文本解析
+4. 卸载时会保留 `argus-agent` 用户，避免影响其他服务

src/metric/client-plugins/all-in-one-full/plugins/argus-agent/check_health.sh

@@ -0,0 +1,69 @@
+#!/bin/bash
+
+# Argus Agent 健康检查脚本
+# 输出 JSON 格式结果
+
+set -e
+
+# 检查 Argus Agent 健康状态
+check_health() {
+    local name="argus-agent"
+    local status="unhealth"
+    local reason=""
+    local install_record="/opt/argus-metric/current/.install_record"
+    
+    # 首先尝试通过安装记录文件检查进程
+    if [[ -f "$install_record" ]]; then
+        # 尝试使用jq解析JSON格式的安装记录文件
+        local pid=""
+        if command -v jq &> /dev/null; then
+            pid=$(jq -r '.components."argus-agent".pid // empty' "$install_record" 2>/dev/null || echo "")
+        else
+            # 如果没有jq，使用简单的文本解析方法
+            pid=$(grep -A 10 '"argus-agent"' "$install_record" | grep '"pid"' | cut -d'"' -f4 | head -1)
+        fi
+        
+        if [[ -n "$pid" && "$pid" =~ ^[0-9]+$ ]]; then
+            if kill -0 "$pid" 2>/dev/null; then
+                # 进程存在且运行正常
+                status="health"
+                reason="进程运行正常 (PID: $pid)"
+                echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+                exit 0
+            else
+                reason="安装记录中的 PID $pid 进程不存在"
+                echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+                exit 1
+            fi
+        else
+            reason="安装记录文件中未找到有效的 argus-agent PID"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    else
+        # 如果安装记录文件不存在，尝试查找 argus-agent 进程
+        local pids=$(pgrep -f "argus-agent" 2>/dev/null || true)
+        if [[ -n "$pids" ]]; then
+            # 取第一个找到的 PID
+            local pid=$(echo "$pids" | head -1)
+            status="health"
+            reason="发现 argus-agent 进程运行 (PID: $pid)，但未找到安装记录"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 0
+        else
+            reason="未找到 argus-agent 进程，且安装记录文件不存在"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    fi
+}
+
+# 主函数
+main() {
+    check_health
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/argus-agent/install.sh

@@ -0,0 +1,289 @@
+#!/bin/bash
+
+set -e
+
+# 颜色定义
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# 日志函数
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+log_warning() {
+    echo -e "${YELLOW}[WARNING]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# 显示帮助信息
+show_help() {
+    echo "Argus Agent 安装脚本"
+    echo
+    echo "用法: $0 [选项]"
+    echo
+    echo "选项:"
+    echo "  --help     显示此帮助信息"
+    echo
+    echo "示例:"
+    echo "  $0          # 安装 Argus Agent"
+    echo
+}
+
+# 解析命令行参数
+INSTALL_DIR=""
+for arg in "$@"; do
+    case $arg in
+        --help|-h)
+            show_help
+            exit 0
+            ;;
+        *)
+            # 如果参数不是以--开头，则认为是安装目录
+            if [[ ! "$arg" =~ ^-- ]]; then
+                INSTALL_DIR="$arg"
+            else
+                log_error "未知参数: $arg"
+                show_help
+                exit 1
+            fi
+            ;;
+    esac
+done
+
+# 检查是否为 root 用户
+check_root() {
+    if [[ $EUID -ne 0 ]]; then
+        log_error "此脚本需要 root 权限运行"
+        log_info "请使用: sudo $0"
+        exit 1
+    fi
+}
+
+# 检查系统要求
+check_system() {
+    log_info "检查系统要求..."
+    
+    # 检查操作系统
+    if [[ ! -f /etc/os-release ]]; then
+        log_error "无法检测操作系统版本"
+        exit 1
+    fi
+    
+    source /etc/os-release
+    log_info "检测到操作系统: $NAME $VERSION"
+    
+    # 检查是否为 Linux 系统
+    if [[ "$ID" != "ubuntu" && "$ID" != "debian" && "$ID" != "centos" && "$ID" != "rhel" && "$ID" != "fedora" ]]; then
+        log_warning "此脚本主要针对常见 Linux 发行版，其他系统可能需要调整"
+    fi
+    
+    # 检查系统架构
+    local arch=$(uname -m)
+    log_info "系统架构: $arch"
+    
+    if [[ "$arch" != "x86_64" && "$arch" != "amd64" ]]; then
+        log_warning "当前架构为 $arch，argus-agent 主要支持 x86_64/amd64"
+    fi
+}
+
+# 停止可能运行的服务
+stop_existing_service() {
+    log_info "检查并停止可能运行的服务..."
+    local pid_file="/var/run/argus-agent.pid"
+
+    if [[ -f "$pid_file" ]]; then
+        local pid=$(cat "$pid_file")
+        if ps -p "$pid" -o comm= | grep -q "^argus-agent$"; then
+            kill "$pid" 2>/dev/null || true
+            sleep 2
+            kill -9 "$pid" 2>/dev/null || true
+            log_success "服务已停止"
+        fi
+        rm -f "$pid_file"
+    fi
+
+    local pids=$(pgrep -x argus-agent 2>/dev/null || true)
+    if [[ -n "$pids" ]]; then
+        for pid in $pids; do kill -9 "$pid" 2>/dev/null || true; done
+    fi
+
+    # 检查僵尸进程
+    local zombies=$(ps -eo pid,stat,comm | grep '[a]rgus-agent' | awk '$2 ~ /Z/ {print $1}')
+    if [[ -n "$zombies" ]]; then
+        for pid in $zombies; do
+            local ppid=$(ps -o ppid= -p $pid)
+            log_warning "检测到僵尸 argus-agent (PID=$pid, PPID=$ppid)，尝试清理"
+            [[ "$ppid" -ne 1 ]] && kill -9 "$ppid" 2>/dev/null || true
+        done
+    fi
+}
+
+
+# 安装 Argus Agent 二进制文件
+install_argus_agent() {
+    log_info "安装 Argus Agent..."
+    local binary_file="bin/argus-agent"
+    local install_dir="/usr/local/bin"
+    local target_file="$install_dir/argus-agent"
+
+    [[ ! -f "$binary_file" ]] && log_error "找不到 Argus Agent 二进制文件: $binary_file" && exit 1
+
+    stop_existing_service
+
+    local timeout=10
+    while [[ $timeout -gt 0 ]]; do
+        remaining_pids=$(pgrep -x argus-agent | grep -vw $$ || true)
+        [[ -z "$remaining_pids" ]] && break
+        if ps -eo pid,stat,comm | grep -E 'argus-agent' | grep -q 'Z'; then
+            log_warning "检测到僵尸 argus-agent，跳过等待"
+            break
+        fi
+        log_warning "等待 argus-agent 完全退出... ($timeout)"
+        sleep 1
+        ((timeout--))
+    done
+
+    cp "$binary_file" "${target_file}.new"
+    chmod +x "${target_file}.new"
+    mv -f "${target_file}.new" "$target_file"
+    log_success "Argus Agent 二进制文件安装完成"
+}
+
+
+# 创建用户和组
+create_user() {
+    log_info "创建 argus-agent 用户..."
+    
+    # 检查用户是否已存在
+    if id "argus-agent" &>/dev/null; then
+        log_info "用户 argus-agent 已存在"
+    else
+        useradd --no-create-home --shell /bin/false argus-agent
+        log_success "用户 argus-agent 创建完成"
+    fi
+}
+
+# 安装配置文件
+install_config() {
+    log_info "安装配置文件..."
+    
+    local config_dir="/etc/argus-agent"
+    
+    # 创建配置目录
+    mkdir -p "$config_dir"
+    
+    # 创建健康检查目录
+    mkdir -p "/var/lib/argus-agent/health"
+    chown argus-agent:argus-agent "/var/lib/argus-agent/health"
+}
+
+# 启动 Argus Agent 服务
+start_argus_agent() {
+    log_info "启动 Argus Agent 服务..."
+    local binary_path="/usr/local/bin/argus-agent"
+    local log_file="/var/log/argus-agent.log"
+    local pid_file="/var/run/argus-agent.pid"
+
+    [[ -f "$pid_file" ]] && rm -f "$pid_file"
+
+    log_info "正在启动 Argus Agent..."
+    setsid "$binary_path" > "$log_file" 2>&1 < /dev/null &
+    local pid=$!
+    echo "$pid" > "$pid_file"
+    sleep 2
+
+    if kill -0 "$pid" 2>/dev/null; then
+        log_success "Argus Agent 服务启动成功 (PID: $pid)"
+    else
+        log_error "Argus Agent 启动失败"
+        [[ -f "$log_file" ]] && tail -n 10 "$log_file"
+        rm -f "$pid_file"
+    fi
+}
+
+
+# 更新安装记录
+update_install_record() {
+    local pid="$1"
+    # 使用传入的安装目录参数，如果没有则使用默认值
+    local install_base_dir="${2:-/opt/argus-metric/current}"
+    local install_record="$install_base_dir/.install_record"
+    
+    # 如果安装记录文件不存在，说明是首次安装，由主安装脚本统一创建
+    if [[ ! -f "$install_record" ]]; then
+        log_info "安装记录文件不存在，将由主安装脚本创建"
+        return 0
+    fi
+    
+    # 如果文件存在，说明是重启场景，只更新 PID 字段
+    if command -v jq &> /dev/null; then
+        # 读取当前 PID
+        local current_pid=$(jq -r '.components."argus-agent".pid // ""' "$install_record" 2>/dev/null)
+        
+        if [[ -z "$current_pid" ]]; then
+            log_warning "无法读取当前 PID，跳过更新"
+            return 1
+        fi
+        
+        # 使用 jq 只更新 pid 字段，保持字符串类型，保留其他字段
+        jq --arg new_pid "$pid" '.components."argus-agent".pid = $new_pid' "$install_record" > "$install_record.tmp" && mv "$install_record.tmp" "$install_record"
+        log_info "PID 已更新: $current_pid -> $pid"
+    else
+        log_warning "jq 命令不可用，无法更新安装记录文件"
+    fi
+}
+
+# 显示安装信息
+show_install_info() {
+    log_success "Argus Agent 安装完成！"
+    echo
+    echo "安装信息:"
+    echo "  二进制文件: /usr/local/bin/argus-agent"
+    echo "  运行用户: argus-agent"
+    echo "  配置目录: /etc/argus-agent/"
+    echo "  健康检查目录: /var/lib/argus-agent/health"
+    echo
+    echo "使用方法:"
+    echo "  手动启动: /usr/local/bin/argus-agent"
+    echo "  后台启动: nohup /usr/local/bin/argus-agent &"
+    echo
+    echo "健康检查:"
+    echo "  ./check_health.sh"
+    echo
+}
+
+# 主函数
+main() {
+    echo "=========================================="
+    echo "    Argus Agent 安装脚本 v1.0"
+    echo "=========================================="
+    echo
+    
+    check_root
+    check_system
+    
+    log_info "开始安装 Argus Agent..."
+    
+    install_argus_agent
+    create_user
+    install_config
+    start_argus_agent
+    
+    show_install_info
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/argus-agent/package.sh

@@ -0,0 +1,87 @@
+#!/bin/bash
+
+set -e
+
+# 颜色定义
+GREEN='\033[0;32m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+# 获取当前目录
+CURRENT_DIR=$(pwd)
+PACKAGE_NAME="argus-agent-$(date +%Y%m%d-%H%M%S)"
+PACKAGE_FILE="${PACKAGE_NAME}.tar.gz"
+
+log_info "开始打包 Argus Agent 安装包..."
+
+# 检查必要文件
+log_info "检查必要文件..."
+
+required_files=(
+    "install.sh"
+    "uninstall.sh"
+    "bin/argus-agent"
+    "check_health.sh"
+)
+
+missing_files=()
+for file in "${required_files[@]}"; do
+    if [[ ! -f "$file" ]]; then
+        missing_files+=("$file")
+    fi
+done
+
+if [[ ${#missing_files[@]} -gt 0 ]]; then
+    echo "缺少以下文件:"
+    for file in "${missing_files[@]}"; do
+        echo "  - $file"
+    done
+    exit 1
+fi
+
+log_success "所有必要文件检查完成"
+
+# 创建临时目录
+TEMP_DIR=$(mktemp -d)
+log_info "创建临时目录: $TEMP_DIR"
+
+# 复制文件到临时目录
+cp -r . "$TEMP_DIR/$PACKAGE_NAME"
+
+# 进入临时目录
+cd "$TEMP_DIR"
+
+# 创建压缩包
+log_info "创建压缩包: $PACKAGE_FILE"
+tar -czf "$PACKAGE_FILE" "$PACKAGE_NAME"
+
+# 移动压缩包到原目录
+mv "$PACKAGE_FILE" "$CURRENT_DIR/"
+
+# 清理临时目录
+rm -rf "$TEMP_DIR"
+
+# 返回原目录
+cd "$CURRENT_DIR"
+
+# 显示结果
+log_success "打包完成！"
+echo
+echo "安装包文件: $PACKAGE_FILE"
+echo "文件大小: $(du -h "$PACKAGE_FILE" | cut -f1)"
+echo
+echo "使用方法:"
+echo "1. 将 $PACKAGE_FILE 传输到目标服务器"
+echo "2. 解压: tar -xzf $PACKAGE_FILE"
+echo "3. 进入目录: cd $PACKAGE_NAME"
+echo "4. 运行安装: sudo ./install.sh"
+echo
+echo "注意: 请确保所有必要文件都存在"

src/metric/client-plugins/all-in-one-full/plugins/argus-agent/uninstall.sh

@@ -0,0 +1,255 @@
+#!/bin/bash
+
+# Argus Agent 卸载脚本
+# 版本: 1.0
+# 作者: AIOps Team
+# 日期: $(date +%Y-%m-%d)
+
+set -e
+
+# 颜色定义
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# 日志函数
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+log_warning() {
+    echo -e "${YELLOW}[WARNING]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# 检查是否为 root 用户
+check_root() {
+    if [[ $EUID -ne 0 ]]; then
+        log_error "此脚本需要 root 权限运行"
+        log_info "请使用: sudo $0"
+        exit 1
+    fi
+}
+
+# 停止运行中的进程
+stop_processes() {
+    log_info "停止 Argus Agent 进程..."
+    
+    local pid_file="/var/run/argus-agent.pid"
+    local stopped=false
+    
+    # 首先尝试通过 PID 文件停止服务
+    if [[ -f "$pid_file" ]]; then
+        local pid=$(cat "$pid_file")
+        if kill -0 "$pid" 2>/dev/null; then
+            log_info "通过 PID 文件停止服务 (PID: $pid)..."
+            kill "$pid"
+            sleep 3
+            
+            # 检查进程是否已停止
+            if kill -0 "$pid" 2>/dev/null; then
+                log_warning "进程未响应，强制终止..."
+                kill -9 "$pid" 2>/dev/null || true
+            fi
+            log_success "Argus Agent 进程已停止"
+            stopped=true
+        else
+            log_warning "PID 文件存在但进程已不存在，清理 PID 文件"
+            rm -f "$pid_file"
+        fi
+    fi
+    
+    # 查找并杀死所有 argus-agent 进程
+    local pids=$(pgrep -f "argus-agent" 2>/dev/null || true)
+    if [[ -n "$pids" ]]; then
+        log_info "发现 argus-agent 进程，正在停止..."
+        for pid in $pids; do
+            log_info "停止进程 PID: $pid"
+            kill "$pid" 2>/dev/null || true
+        done
+        sleep 2
+        
+        # 检查是否还有进程在运行，如果有则强制终止
+        local remaining_pids=$(pgrep -f "argus-agent" 2>/dev/null || true)
+        if [[ -n "$remaining_pids" ]]; then
+            log_warning "进程未响应，强制终止..."
+            for pid in $remaining_pids; do
+                log_info "强制终止进程 PID: $pid"
+                kill -9 "$pid" 2>/dev/null || true
+            done
+            sleep 1
+        fi
+        
+        # 最终检查
+        if pgrep -f "argus-agent" > /dev/null; then
+            log_error "无法停止所有 argus-agent 进程"
+        else
+            log_success "所有 Argus Agent 进程已停止"
+            stopped=true
+        fi
+    else
+        log_info "Argus Agent 进程未运行"
+    fi
+    
+    # 清理 PID 文件
+    rm -f "$pid_file"
+    
+    if [[ "$stopped" == "false" ]]; then
+        log_warning "未发现需要停止的 Argus Agent 进程"
+    fi
+}
+
+# 删除二进制文件
+remove_binary() {
+    log_info "删除 Argus Agent 二进制文件..."
+    
+    local binary_files=(
+        "/usr/local/bin/argus-agent"
+    )
+    
+    local deleted=false
+    for binary_file in "${binary_files[@]}"; do
+        if [[ -f "$binary_file" ]]; then
+            rm -f "$binary_file"
+            log_success "二进制文件已删除: $binary_file"
+            deleted=true
+        fi
+    done
+    
+    if [[ "$deleted" == "false" ]]; then
+        log_info "二进制文件不存在"
+    fi
+}
+
+# 删除配置文件
+remove_config() {
+    log_info "删除配置文件..."
+    
+    local config_dir="/etc/argus-agent"
+    
+    if [[ -d "$config_dir" ]]; then
+        rm -rf "$config_dir"
+        log_success "配置目录已删除"
+    else
+        log_info "配置目录不存在"
+    fi
+}
+
+# 删除数据目录
+remove_data_dir() {
+    log_info "删除数据目录..."
+    
+    local data_dir="/var/lib/argus-agent"
+    
+    if [[ -d "$data_dir" ]]; then
+        rm -rf "$data_dir"
+        log_success "数据目录已删除"
+    else
+        log_info "数据目录不存在"
+    fi
+}
+
+# 检查用户状态（可选）
+check_user_status() {
+    log_info "检查 argus-agent 用户状态..."
+    
+    if id "argus-agent" &>/dev/null; then
+        log_info "检测到 argus-agent 用户存在"
+        log_warning "argus-agent 是系统用户，可能被其他服务使用"
+        log_info "为了系统稳定性，将保留 argus-agent 用户"
+        log_info "如需手动删除，请运行: sudo userdel argus-agent"
+    else
+        log_info "argus-agent 用户不存在"
+    fi
+}
+
+# 清理日志文件
+cleanup_logs() {
+    log_info "清理日志文件..."
+    
+    # 删除安装脚本创建的日志文件
+    rm -f /var/log/argus-agent.log
+    
+    log_success "日志文件已清理"
+}
+
+# 清理安装记录
+cleanup_install_record() {
+    log_info "清理安装记录..."
+    
+    local install_record="/opt/argus-metric/current/.install_record"
+    
+    if [[ -f "$install_record" ]]; then
+        if command -v jq &> /dev/null; then
+            # 使用 jq 删除 argus-agent 记录
+            jq 'del(.components."argus-agent")' "$install_record" > "$install_record.tmp" && mv "$install_record.tmp" "$install_record"
+            log_success "安装记录已更新"
+        else
+            log_warning "jq 命令不可用，无法清理安装记录"
+        fi
+    else
+        log_info "安装记录文件不存在"
+    fi
+}
+
+# 显示卸载信息
+show_uninstall_info() {
+    log_success "Argus Agent 卸载完成！"
+    echo
+    echo "已删除的内容:"
+    echo "  - 二进制文件: /usr/local/bin/argus-agent"
+    echo "  - 配置目录: /etc/argus-agent"
+    echo "  - 数据目录: /var/lib/argus-agent"
+    echo "  - 相关日志文件"
+    echo
+    echo "注意:"
+    echo "  - argus-agent 用户已保留（系统用户，可能被其他服务使用）"
+    echo "  - 如需完全清理，请手动检查并删除相关文件"
+    echo
+}
+
+# 主函数
+main() {
+    echo "=========================================="
+    echo "    Argus Agent 卸载脚本 v1.0"
+    echo "=========================================="
+    echo
+    
+    check_root
+    
+    log_warning "此操作将完全卸载 Argus Agent"
+    read -p "确认继续？(y/N): " confirm
+    
+    if [[ "$confirm" != "y" && "$confirm" != "Y" ]]; then
+        log_info "取消卸载操作"
+        exit 0
+    fi
+    
+    log_info "开始卸载 Argus Agent..."
+    
+    stop_processes
+    remove_binary
+    remove_config
+    remove_data_dir
+    cleanup_logs
+    cleanup_install_record
+    
+    # 检查用户状态
+    check_user_status
+    
+    show_uninstall_info
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/check_health.sh

@@ -0,0 +1,55 @@
+#!/bin/bash
+
+# DCGM Exporter 健康检查脚本
+# 输出 JSON 格式结果
+
+set -e
+
+# 检查 DCGM Exporter 健康状态
+check_health() {
+    local url="http://localhost:9400"
+    local metrics_url="$url/metrics"
+    local name="dcgm-exporter"
+    local status="unhealth"
+    local reason=""
+    
+    # 检查 curl 是否可用
+    if ! command -v curl &> /dev/null; then
+        reason="curl 命令不可用，无法进行健康检查"
+        echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+        exit 1
+    fi
+    
+    # 测试根路径连接
+    local http_code=$(curl -s -o /dev/null -w "%{http_code}" "$url" 2>/dev/null || echo "000")
+    
+    if [[ "$http_code" == "200" ]]; then
+        # 测试 metrics 端点
+        local metrics_code=$(curl -s -o /dev/null -w "%{http_code}" "$metrics_url" 2>/dev/null || echo "000")
+        
+        if [[ "$metrics_code" == "200" ]]; then
+            status="health"
+            reason="success"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 0
+        else
+            reason="Metrics 端点异常 (HTTP $metrics_code)"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    else
+        reason="HTTP 服务异常 (HTTP $http_code)，请检查 DCGM Exporter 是否正在运行在端口 9400"
+        echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+        exit 1
+    fi
+}
+
+# 主函数
+main() {
+    check_health
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/config/default-counters.csv

@@ -0,0 +1,77 @@
+# Format
+# If line starts with a '#' it is considered a comment
+# DCGM FIELD, Prometheus metric type, help message
+
+# Clocks
+DCGM_FI_DEV_SM_CLOCK,  gauge, SM clock frequency (in MHz).
+DCGM_FI_DEV_MEM_CLOCK, gauge, Memory clock frequency (in MHz).
+# DCGM_EXP_CLOCK_EVENTS_COUNT, gauge, Count of clock events within the user-specified time window (see clock-events-count-window-size param).
+
+# Temperature
+DCGM_FI_DEV_MEMORY_TEMP, gauge, Memory temperature (in C).
+DCGM_FI_DEV_GPU_TEMP,    gauge, GPU temperature (in C).
+
+# Power
+DCGM_FI_DEV_POWER_USAGE,              gauge, Power draw (in W).
+DCGM_FI_DEV_TOTAL_ENERGY_CONSUMPTION, counter, Total energy consumption since boot (in mJ).
+
+# PCIE
+DCGM_FI_PROF_PCIE_TX_BYTES,  counter, Total number of bytes transmitted through PCIe TX via NVML.
+DCGM_FI_PROF_PCIE_RX_BYTES,  counter, Total number of bytes received through PCIe RX via NVML.
+DCGM_FI_DEV_PCIE_REPLAY_COUNTER, counter, Total number of PCIe retries.
+
+# Utilization (the sample period varies depending on the product)
+DCGM_FI_DEV_GPU_UTIL,      gauge, GPU utilization (in %).
+DCGM_FI_DEV_MEM_COPY_UTIL, gauge, Memory utilization (in %).
+DCGM_FI_DEV_ENC_UTIL,      gauge, Encoder utilization (in %).
+DCGM_FI_DEV_DEC_UTIL ,     gauge, Decoder utilization (in %).
+
+# Errors and violations
+DCGM_FI_DEV_XID_ERRORS,              gauge,   Value of the last XID error encountered.
+# DCGM_FI_DEV_POWER_VIOLATION,       counter, Throttling duration due to power constraints (in us).
+# DCGM_FI_DEV_THERMAL_VIOLATION,     counter, Throttling duration due to thermal constraints (in us).
+# DCGM_FI_DEV_SYNC_BOOST_VIOLATION,  counter, Throttling duration due to sync-boost constraints (in us).
+# DCGM_FI_DEV_BOARD_LIMIT_VIOLATION, counter, Throttling duration due to board limit constraints (in us).
+# DCGM_FI_DEV_LOW_UTIL_VIOLATION,    counter, Throttling duration due to low utilization (in us).
+# DCGM_FI_DEV_RELIABILITY_VIOLATION, counter, Throttling duration due to reliability constraints (in us).
+# DCGM_EXP_XID_ERRORS_COUNT,         gauge,   Count of XID Errors within user-specified time window (see xid-count-window-size param).
+# Memory usage
+DCGM_FI_DEV_FB_FREE, gauge, Frame buffer memory free (in MB).
+DCGM_FI_DEV_FB_USED, gauge, Frame buffer memory used (in MB).
+
+# ECC
+# DCGM_FI_DEV_ECC_SBE_VOL_TOTAL, counter, Total number of single-bit volatile ECC errors.
+# DCGM_FI_DEV_ECC_DBE_VOL_TOTAL, counter, Total number of double-bit volatile ECC errors.
+# DCGM_FI_DEV_ECC_SBE_AGG_TOTAL, counter, Total number of single-bit persistent ECC errors.
+# DCGM_FI_DEV_ECC_DBE_AGG_TOTAL, counter, Total number of double-bit persistent ECC errors.
+
+# Retired pages
+# DCGM_FI_DEV_RETIRED_SBE,     counter, Total number of retired pages due to single-bit errors.
+# DCGM_FI_DEV_RETIRED_DBE,     counter, Total number of retired pages due to double-bit errors.
+# DCGM_FI_DEV_RETIRED_PENDING, counter, Total number of pages pending retirement.
+
+# NVLink
+# DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_TOTAL, counter, Total number of NVLink flow-control CRC errors.
+# DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_TOTAL, counter, Total number of NVLink data CRC errors.
+# DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_TOTAL,   counter, Total number of NVLink retries.
+# DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_TOTAL, counter, Total number of NVLink recovery errors.
+DCGM_FI_DEV_NVLINK_BANDWIDTH_TOTAL,            counter, Total number of NVLink bandwidth counters for all lanes
+
+# VGPU License status
+DCGM_FI_DEV_VGPU_LICENSE_STATUS, gauge, vGPU License status
+
+# Remapped rows
+DCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS, counter, Number of remapped rows for uncorrectable errors
+DCGM_FI_DEV_CORRECTABLE_REMAPPED_ROWS,   counter, Number of remapped rows for correctable errors
+DCGM_FI_DEV_ROW_REMAP_FAILURE,           gauge,   Whether remapping of rows has failed
+
+# Static configuration information. These appear as labels on the other metrics
+DCGM_FI_DRIVER_VERSION,        label, Driver Version
+# DCGM_FI_NVML_VERSION,          label, NVML Version
+# DCGM_FI_DEV_BRAND,             label, Device Brand
+# DCGM_FI_DEV_SERIAL,            label, Device Serial Number
+# DCGM_FI_DEV_OEM_INFOROM_VER,   label, OEM inforom version
+# DCGM_FI_DEV_ECC_INFOROM_VER,   label, ECC inforom version
+# DCGM_FI_DEV_POWER_INFOROM_VER, label, Power management object inforom version
+# DCGM_FI_DEV_INFOROM_IMAGE_VER, label, Inforom image version
+# DCGM_FI_DEV_VBIOS_VERSION,     label, VBIOS version of the device

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/install.sh

@@ -0,0 +1,365 @@
+#!/bin/bash
+
+set -e
+
+# 颜色定义
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# 日志函数
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+log_warning() {
+    echo -e "${YELLOW}[WARNING]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# 更新安装记录
+update_install_record() {
+    local pid="$1"
+    # 使用传入的安装目录参数，如果没有则使用默认值
+    local install_base_dir="${2:-/opt/argus-metric/current}"
+    local install_record="$install_base_dir/.install_record"
+    
+    # 如果安装记录文件不存在，说明是首次安装，由主安装脚本统一创建
+    if [[ ! -f "$install_record" ]]; then
+        log_info "安装记录文件不存在，将由主安装脚本创建"
+        return 0
+    fi
+    
+    # 如果文件存在，说明是重启场景，只更新 PID 字段
+    if command -v jq &> /dev/null; then
+        # 读取当前 PID
+        local current_pid=$(jq -r '.components."dcgm-exporter".pid // ""' "$install_record" 2>/dev/null)
+        
+        if [[ -z "$current_pid" ]]; then
+            log_warning "无法读取当前 PID，跳过更新"
+            return 1
+        fi
+        
+        # 使用 jq 只更新 pid 字段，保持字符串类型，保留其他字段
+        jq --arg new_pid "$pid" '.components."dcgm-exporter".pid = $new_pid' "$install_record" > "$install_record.tmp" && mv "$install_record.tmp" "$install_record"
+        log_info "PID 已更新: $current_pid -> $pid"
+    else
+        log_warning "jq 命令不可用，无法更新安装记录文件"
+    fi
+}
+
+# 显示帮助信息
+show_help() {
+    echo "DCGM Exporter 安装脚本"
+    echo
+    echo "用法: $0 [选项]"
+    echo
+    echo "选项:"
+    echo "  --help     显示此帮助信息"
+    echo
+    echo "示例:"
+    echo "  $0          # 安装 DCGM Exporter"
+    echo
+}
+
+# 解析命令行参数
+INSTALL_DIR=""
+for arg in "$@"; do
+    case $arg in
+        --help|-h)
+            show_help
+            exit 0
+            ;;
+        *)
+            # 如果参数不是以--开头，则认为是安装目录
+            if [[ ! "$arg" =~ ^-- ]]; then
+                INSTALL_DIR="$arg"
+            else
+                log_error "未知参数: $arg"
+                show_help
+                exit 1
+            fi
+            ;;
+    esac
+done
+
+# 检查是否为 root 用户
+check_root() {
+    if [[ $EUID -ne 0 ]]; then
+        log_error "此脚本需要 root 权限运行"
+        log_info "请使用: sudo $0"
+        exit 1
+    fi
+}
+
+# 检查系统要求
+check_system() {
+    log_info "检查系统要求..."
+    
+    # 检查操作系统
+    if [[ ! -f /etc/os-release ]]; then
+        log_error "无法检测操作系统版本"
+        exit 1
+    fi
+    
+    source /etc/os-release
+    log_info "检测到操作系统: $NAME $VERSION"
+    
+    # 检查是否为 Ubuntu/Debian
+    if [[ "$ID" != "ubuntu" && "$ID" != "debian" ]]; then
+        log_warning "此脚本主要针对 Ubuntu/Debian 系统，其他系统可能需要调整"
+    fi
+    
+    # 检查 NVIDIA GPU
+    if ! command -v nvidia-smi &> /dev/null; then
+        log_warning "未检测到 nvidia-smi，请确保已安装 NVIDIA 驱动"
+    else
+        log_success "检测到 NVIDIA GPU"
+        nvidia-smi --query-gpu=name --format=csv,noheader,nounits | head -1
+    fi
+}
+
+# 安装 DCGM 依赖
+install_dcgm_dependency() {
+    log_info "安装 DCGM 依赖..."
+    
+    local deb_file="bin/datacenter-gpu-manager_3.3.9_amd64.deb"
+    
+    if [[ ! -f "$deb_file" ]]; then
+        log_error "找不到 DCGM 依赖文件: $deb_file"
+        exit 1
+    fi
+    
+    # 安装 deb 包
+    dpkg -i "$deb_file" || {
+        log_warning "dpkg 安装失败，尝试使用 apt 修复依赖..."
+        apt-get update
+        apt-get install -f -y
+        dpkg -i "$deb_file"
+    }
+    
+    log_success "DCGM 依赖安装完成"
+}
+
+# 检查 DCGM 服务状态
+check_dcgm_service() {
+    log_info "检查 DCGM 服务状态..."
+    
+    # 检查 DCGM 服务是否在运行
+    if systemctl is-active --quiet dcgm 2>/dev/null; then
+        log_success "DCGM 服务已在运行"
+    elif pgrep -f nv-hostengine > /dev/null; then
+        log_success "nv-hostengine 进程已在运行"
+    else
+        log_warning "DCGM 服务未运行，需要手动启动"
+        log_info "启动 DCGM 服务的方法:"
+        log_info "  1. 使用 systemd: sudo systemctl start dcgm"
+        log_info "  2. 手动启动: nohup nv-hostengine > /var/log/nv-hostengine.log 2>&1 &"
+    fi
+    
+    # 测试 DCGM 连接
+    if systemctl is-active --quiet dcgm 2>/dev/null || pgrep -f nv-hostengine > /dev/null; then
+        log_info "测试 DCGM 连接..."
+        if dcgmi discovery -l > /dev/null 2>&1; then
+            log_success "DCGM 连接测试成功"
+        else
+            log_warning "DCGM 连接测试失败，请检查服务状态"
+        fi
+    fi
+}
+
+# 停止可能运行的服务
+stop_existing_service() {
+    log_info "检查并停止可能运行的服务..."
+    
+    local pid_file="/var/run/dcgm-exporter.pid"
+    
+    # 检查并停止通过 PID 文件管理的服务
+    if [[ -f "$pid_file" ]]; then
+        local pid=$(cat "$pid_file")
+        if kill -0 "$pid" 2>/dev/null; then
+            log_info "发现正在运行的 DCGM Exporter 服务 (PID: $pid)，正在停止..."
+            kill "$pid" > /dev/null 2>&1 || true
+            sleep 2
+            if kill -0 "$pid" 2>/dev/null; then
+                log_warning "进程未响应，强制终止..."
+                kill -9 "$pid" > /dev/null 2>&1 || true
+            fi
+            rm -f "$pid_file"
+            log_success "服务已停止"
+        else
+            log_warning "发现过期的 PID 文件，正在清理..."
+            rm -f "$pid_file"
+        fi
+    fi
+    
+    # 查找并停止所有 dcgm-exporter 进程（排除脚本自身）
+    local exporter_bin="/usr/local/bin/dcgm-exporter"
+    local pids=$(pgrep -f "$exporter_bin")
+
+    if [[ -n "$pids" ]]; then
+        log_info "发现其他 dcgm-exporter 进程，正在停止..."
+        for pid in $pids; do
+            if [[ "$pid" != "$$" ]]; then
+                kill "$pid" > /dev/null 2>&1 || true
+                sleep 1
+                if kill -0 "$pid" 2>/dev/null; then
+                    log_warning "进程 $pid 未响应，强制终止..."
+                    kill -9 "$pid" > /dev/null 2>&1 || true
+                fi
+            fi
+        done
+        log_success "所有 dcgm-exporter 进程已停止"
+    fi
+}
+
+# 安装 DCGM Exporter 二进制文件
+install_dcgm_exporter() {
+    log_info "安装 DCGM Exporter..."
+    
+    local binary_file="bin/dcgm-exporter"
+    local install_dir="/usr/local/bin"
+    
+    if [[ ! -f "$binary_file" ]]; then
+        log_error "找不到 DCGM Exporter 二进制文件: $binary_file"
+        exit 1
+    fi
+    
+    # 停止可能运行的服务
+    stop_existing_service
+    
+    # 复制二进制文件
+    cp "$binary_file" "$install_dir/"
+    chmod +x "$install_dir/dcgm-exporter"
+    
+    log_success "DCGM Exporter 二进制文件安装完成"
+}
+
+# 安装配置文件
+install_config() {
+    log_info "安装配置文件..."
+    
+    local config_dir="/etc/dcgm-exporter"
+    local config_file="config/default-counters.csv"
+    
+    # 创建配置目录
+    mkdir -p "$config_dir"
+    
+    if [[ -f "$config_file" ]]; then
+        cp "$config_file" "$config_dir/"
+        log_success "配置文件安装完成"
+    else
+        log_warning "未找到配置文件，使用默认配置"
+    fi
+}
+
+# 启动 DCGM Exporter 服务
+start_dcgm_exporter() {
+    log_info "启动 DCGM Exporter 服务..."
+    
+    local binary_path="/usr/local/bin/dcgm-exporter"
+    local log_file="/var/log/dcgm-exporter.log"
+    local pid_file="/var/run/dcgm-exporter.pid"
+    
+    # 检查服务是否已经在运行
+    if [[ -f "$pid_file" ]]; then
+        local pid=$(cat "$pid_file")
+        if kill -0 "$pid" 2>/dev/null; then
+            log_info "DCGM Exporter 服务已在运行 (PID: $pid)"
+            return 0
+        else
+            log_warning "发现过期的 PID 文件，正在清理..."
+            rm -f "$pid_file"
+        fi
+    fi
+    
+    # 检查端口是否被占用
+    if netstat -tuln 2>/dev/null | grep -q ":9400 "; then
+        log_warning "端口 9400 已被占用，请检查是否有其他服务在运行"
+        return 1
+    fi
+    
+    # 启动服务
+    log_info "正在启动 DCGM Exporter..."
+    nohup "$binary_path" --address=:9400 > "$log_file" 2>&1 &
+    local pid=$!
+    
+    # 保存 PID
+    echo "$pid" > "$pid_file"
+    
+    # 等待服务启动
+    sleep 2
+    
+    # 检查服务是否成功启动
+    if kill -0 "$pid" 2>/dev/null; then
+        log_success "DCGM Exporter 服务启动成功 (PID: $pid)"
+        log_info "日志文件: $log_file"
+        log_info "PID 文件: $pid_file"
+        
+        # 更新安装记录
+        update_install_record "$pid" "$INSTALL_DIR"
+    else
+        log_error "DCGM Exporter 服务启动失败"
+        rm -f "$pid_file"
+        return 1
+    fi
+}
+
+
+
+# 显示安装信息
+show_install_info() {
+    log_success "DCGM Exporter 安装完成！"
+    echo
+    echo "安装信息:"
+    echo "  二进制文件: /usr/local/bin/dcgm-exporter"
+    echo "  配置文件: /etc/dcgm-exporter/default-counters.csv"
+    echo "  默认端口: 9400"
+    echo
+    echo "使用方法:"
+    echo "  1. 启动 DCGM 服务:"
+    echo "     sudo systemctl start dcgm"
+    echo "     或: nohup nv-hostengine > /var/log/nv-hostengine.log 2>&1 &"
+    echo "  2. 启动 DCGM Exporter:"
+    echo "     /usr/local/bin/dcgm-exporter --address=:9400"
+    echo "     或: nohup /usr/local/bin/dcgm-exporter --address=:9400 &"
+    echo
+    echo "测试连接:"
+    echo "  curl http://localhost:9400/metrics"
+    echo
+}
+
+# 主函数
+main() {
+    echo "=========================================="
+    echo "    DCGM Exporter 安装脚本 v1.0"
+    echo "=========================================="
+    echo
+    
+    check_root
+    check_system
+    
+    log_info "开始安装 DCGM Exporter..."
+    
+    install_dcgm_dependency
+    check_dcgm_service
+    install_dcgm_exporter
+    install_config
+    start_dcgm_exporter
+    
+    show_install_info
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/package.sh

@@ -0,0 +1,88 @@
+#!/bin/bash
+
+set -e
+
+# 颜色定义
+GREEN='\033[0;32m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+# 获取当前目录
+CURRENT_DIR=$(pwd)
+PACKAGE_NAME="dcgm-exporter-$(date +%Y%m%d-%H%M%S)"
+PACKAGE_FILE="${PACKAGE_NAME}.tar.gz"
+
+log_info "开始打包 DCGM Exporter 安装包..."
+
+# 检查必要文件
+log_info "检查必要文件..."
+
+required_files=(
+    "install.sh"
+    "uninstall.sh"
+    "bin/dcgm-exporter"
+    "bin/datacenter-gpu-manager_3.3.9_amd64.deb"
+    "check_health.sh"
+)
+
+missing_files=()
+for file in "${required_files[@]}"; do
+    if [[ ! -f "$file" ]]; then
+        missing_files+=("$file")
+    fi
+done
+
+if [[ ${#missing_files[@]} -gt 0 ]]; then
+    echo "缺少以下文件:"
+    for file in "${missing_files[@]}"; do
+        echo "  - $file"
+    done
+    exit 1
+fi
+
+log_success "所有必要文件检查完成"
+
+# 创建临时目录
+TEMP_DIR=$(mktemp -d)
+log_info "创建临时目录: $TEMP_DIR"
+
+# 复制文件到临时目录
+cp -r . "$TEMP_DIR/$PACKAGE_NAME"
+
+# 进入临时目录
+cd "$TEMP_DIR"
+
+# 创建压缩包
+log_info "创建压缩包: $PACKAGE_FILE"
+tar -czf "$PACKAGE_FILE" "$PACKAGE_NAME"
+
+# 移动压缩包到原目录
+mv "$PACKAGE_FILE" "$CURRENT_DIR/"
+
+# 清理临时目录
+rm -rf "$TEMP_DIR"
+
+# 返回原目录
+cd "$CURRENT_DIR"
+
+# 显示结果
+log_success "打包完成！"
+echo
+echo "安装包文件: $PACKAGE_FILE"
+echo "文件大小: $(du -h "$PACKAGE_FILE" | cut -f1)"
+echo
+echo "使用方法:"
+echo "1. 将 $PACKAGE_FILE 传输到目标服务器"
+echo "2. 解压: tar -xzf $PACKAGE_FILE"
+echo "3. 进入目录: cd $PACKAGE_NAME"
+echo "4. 运行安装: sudo ./install.sh"
+echo
+echo "注意: 请确保 config/default-counters.csv 文件存在"

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/uninstall.sh

@@ -0,0 +1,216 @@
+#!/bin/bash
+
+# DCGM Exporter 卸载脚本
+# 版本: 1.0
+# 作者: AIOps Team
+# 日期: $(date +%Y-%m-%d)
+
+set -e
+
+# 颜色定义
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+# 日志函数
+log_info() {
+    echo -e "${BLUE}[INFO]${NC} $1"
+}
+
+log_success() {
+    echo -e "${GREEN}[SUCCESS]${NC} $1"
+}
+
+log_warning() {
+    echo -e "${YELLOW}[WARNING]${NC} $1"
+}
+
+log_error() {
+    echo -e "${RED}[ERROR]${NC} $1"
+}
+
+# 检查是否为 root 用户
+check_root() {
+    if [[ $EUID -ne 0 ]]; then
+        log_error "此脚本需要 root 权限运行"
+        log_info "请使用: sudo $0"
+        exit 1
+    fi
+}
+
+# 停止运行中的进程
+stop_processes() {
+    log_info "停止 DCGM Exporter 进程..."
+    
+    local pid_file="/var/run/dcgm-exporter.pid"
+    local stopped=false
+    
+    # 首先尝试通过 PID 文件停止服务
+    if [[ -f "$pid_file" ]]; then
+        local pid=$(cat "$pid_file")
+        if kill -0 "$pid" 2>/dev/null; then
+            log_info "通过 PID 文件停止服务 (PID: $pid)..."
+            kill "$pid"
+            sleep 3
+            
+            # 检查进程是否已停止
+            if kill -0 "$pid" 2>/dev/null; then
+                log_warning "进程未响应，强制终止..."
+                kill -9 "$pid" 2>/dev/null || true
+            fi
+            log_success "DCGM Exporter 进程已停止"
+            stopped=true
+        else
+            log_warning "PID 文件存在但进程已不存在，清理 PID 文件"
+            rm -f "$pid_file"
+        fi
+    fi
+    
+    # 查找并杀死所有 dcgm-exporter 进程
+    local pids=$(pgrep -f "dcgm-exporter" 2>/dev/null || true)
+    if [[ -n "$pids" ]]; then
+        log_info "发现 dcgm-exporter 进程，正在停止..."
+        for pid in $pids; do
+            log_info "停止进程 PID: $pid"
+            kill "$pid" 2>/dev/null || true
+        done
+        sleep 2
+        
+        # 检查是否还有进程在运行，如果有则强制终止
+        local remaining_pids=$(pgrep -f "dcgm-exporter" 2>/dev/null || true)
+        if [[ -n "$remaining_pids" ]]; then
+            log_warning "进程未响应，强制终止..."
+            for pid in $remaining_pids; do
+                log_info "强制终止进程 PID: $pid"
+                kill -9 "$pid" 2>/dev/null || true
+            done
+            sleep 1
+        fi
+        
+        # 最终检查
+        if pgrep -f "dcgm-exporter" > /dev/null; then
+            log_error "无法停止所有 dcgm-exporter 进程"
+        else
+            log_success "所有 DCGM Exporter 进程已停止"
+            stopped=true
+        fi
+    else
+        log_info "DCGM Exporter 进程未运行"
+    fi
+    
+    # 清理 PID 文件
+    rm -f "$pid_file"
+    
+    if [[ "$stopped" == "false" ]]; then
+        log_warning "未发现需要停止的 DCGM Exporter 进程"
+    fi
+}
+
+# 删除二进制文件
+remove_binary() {
+    log_info "删除 DCGM Exporter 二进制文件..."
+    
+    local binary_file="/usr/local/bin/dcgm-exporter"
+    
+    if [[ -f "$binary_file" ]]; then
+        rm -f "$binary_file"
+        log_success "二进制文件已删除"
+    else
+        log_info "二进制文件不存在"
+    fi
+}
+
+# 删除配置文件
+remove_config() {
+    log_info "删除配置文件..."
+    
+    local config_dir="/etc/dcgm-exporter"
+    
+    if [[ -d "$config_dir" ]]; then
+        rm -rf "$config_dir"
+        log_success "配置目录已删除"
+    else
+        log_info "配置目录不存在"
+    fi
+}
+
+# 卸载 DCGM 依赖（可选）
+remove_dcgm_dependency() {
+    log_info "检查 DCGM 依赖状态..."
+    
+    # 检查是否安装了 DCGM 包
+    if dpkg -l | grep -q datacenter-gpu-manager; then
+        log_info "检测到 DCGM 依赖包已安装"
+        log_warning "DCGM 是系统级依赖，可能被其他应用程序使用"
+        log_info "为了系统稳定性，将保留 DCGM 依赖包"
+        log_info "如需手动卸载，请运行: sudo apt-get remove --purge datacenter-gpu-manager"
+    else
+        log_info "DCGM 依赖包未安装"
+    fi
+}
+
+# 清理日志文件
+cleanup_logs() {
+    log_info "清理日志文件..."
+    
+    # 清理 journal 日志
+    journalctl --vacuum-time=1s --quiet || true
+    
+    # 删除可能的日志文件
+    rm -f /var/log/nv-hostengine.log
+    rm -f /var/log/dcgm-exporter.log
+    
+    log_success "日志文件已清理"
+}
+
+# 显示卸载信息
+show_uninstall_info() {
+    log_success "DCGM Exporter 卸载完成！"
+    echo
+    echo "已删除的内容:"
+    echo "  - 二进制文件: /usr/local/bin/dcgm-exporter"
+    echo "  - 配置目录: /etc/dcgm-exporter"
+    echo "  - 相关日志文件"
+    echo
+    echo "注意:"
+    echo "  - DCGM 依赖包可能仍然存在"
+    echo "  - 如需完全清理，请手动检查并删除相关文件"
+    echo
+}
+
+# 主函数
+main() {
+    echo "=========================================="
+    echo "    DCGM Exporter 卸载脚本 v1.0"
+    echo "=========================================="
+    echo
+    
+    check_root
+    
+    log_warning "此操作将完全卸载 DCGM Exporter"
+    read -p "确认继续？(y/N): " confirm
+    
+    if [[ "$confirm" != "y" && "$confirm" != "Y" ]]; then
+        log_info "取消卸载操作"
+        exit 0
+    fi
+    
+    log_info "开始卸载 DCGM Exporter..."
+    
+    stop_processes
+    remove_binary
+    remove_config
+    cleanup_logs
+    
+    # 询问是否卸载 DCGM 依赖
+    remove_dcgm_dependency
+    
+    show_uninstall_info
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/README.md

@@ -0,0 +1,181 @@
+# Fluent Bit 安装包
+
+这是一个 Fluent Bit 的自动化安装包，提供了完整的安装、卸载和健康检查功能。
+
+## 目录结构
+
+```
+fluent-bit-installer/
+├── install.sh                    # 安装脚本
+├── uninstall.sh                  # 卸载脚本
+├── package.sh                    # 打包脚本
+├── check_health.sh               # 健康检查脚本
+├── bin/
+│   └── fluent-bit_3.1.9_amd64.deb  # Fluent Bit 安装包
+└── config/
+    ├── fluent-bit.conf           # 主配置文件
+    ├── inject_labels.lua         # Lua 脚本
+    ├── parsers.conf              # 解析器配置
+    ├── inputs.d/                 # 输入配置目录
+    │   ├── 10-train.conf
+    │   └── 20-infer.conf
+    └── outputs.d/                # 输出配置目录
+        └── 10-es.conf
+```
+
+## 功能特性
+
+- **自动化安装**: 一键安装 Fluent Bit 及其依赖
+- **配置管理**: 自动部署预配置的配置文件
+- **服务管理**: 自动启动和停止 Fluent Bit 服务
+- **健康检查**: 提供 JSON 格式的健康状态检查
+- **完整卸载**: 彻底清理所有相关文件和配置
+- **用户管理**: 自动创建专用的 fluent-bit 用户
+
+## 使用方法
+
+### 1. 打包安装包
+
+```bash
+./package.sh
+```
+
+这将创建一个带时间戳的压缩包，例如：`fluent-bit-installer-20250924-160954.tar.gz`
+
+### 2. 安装 Fluent Bit
+
+```bash
+# 解压安装包
+tar -xzf fluent-bit-installer-*.tar.gz
+cd fluent-bit-installer-*
+
+# 运行安装脚本（需要 root 权限）
+sudo ./install.sh
+```
+
+### 3. 健康检查
+
+```bash
+./check_health.sh
+```
+
+输出示例：
+```json
+{"name": "fluent-bit", "status": "health", "reason": "success"}
+```
+
+### 4. 卸载 Fluent Bit
+
+```bash
+sudo ./uninstall.sh
+```
+
+## 安装后的文件位置
+
+- **二进制文件**: `/opt/fluent-bit/bin/fluent-bit`
+- **配置文件**: `/etc/fluent-bit/`
+- **日志文件**: `/var/log/fluent-bit/`
+- **缓冲区目录**: `/var/lib/fluent-bit/buffers/`
+- **运行用户**: `fluent-bit`
+- **HTTP 端口**: `2020`
+
+## 配置说明
+
+### 主配置文件
+
+主配置文件位于 `/etc/fluent-bit/fluent-bit.conf`，包含以下主要部分：
+
+- **SERVICE**: 服务配置，包括 HTTP 服务器设置
+- **INPUT**: 输入配置，通过 `inputs.d/` 目录管理
+- **FILTER**: 过滤器配置，包括解析器和标签注入
+- **OUTPUT**: 输出配置，通过 `outputs.d/` 目录管理
+
+### 输入配置
+
+- `10-train.conf`: 训练日志输入配置
+- `20-infer.conf`: 推理日志输入配置
+
+### 输出配置
+
+- `10-es.conf`: Elasticsearch 输出配置
+
+## 服务管理
+
+### 手动启动
+
+```bash
+/opt/fluent-bit/bin/fluent-bit --config=/etc/fluent-bit/fluent-bit.conf
+```
+
+### 后台启动
+
+```bash
+nohup /opt/fluent-bit/bin/fluent-bit --config=/etc/fluent-bit/fluent-bit.conf &
+```
+
+### 检查服务状态
+
+```bash
+# 检查进程
+ps aux | grep fluent-bit
+
+# 检查端口
+netstat -tuln | grep 2020
+
+# 检查日志
+tail -f /var/log/fluent-bit/fluent-bit.log
+```
+
+## API 接口
+
+Fluent Bit 提供 HTTP API 用于监控和管理：
+
+- **根路径**: `http://localhost:2020`
+- **状态接口**: `http://localhost:2020/api/v1/status`
+- **指标接口**: `http://localhost:2020/api/v1/metrics`
+
+## 故障排除
+
+### 常见问题
+
+1. **端口被占用**
+   - 检查端口 2020 是否被其他服务占用
+   - 修改配置文件中的端口设置
+
+2. **权限问题**
+   - 确保 fluent-bit 用户有足够的权限访问日志文件
+   - 检查目录权限设置
+
+3. **配置文件错误**
+   - 检查配置文件语法
+   - 查看日志文件中的错误信息
+
+### 日志查看
+
+```bash
+# 查看服务日志
+tail -f /var/log/fluent-bit/fluent-bit.log
+
+# 查看系统日志
+journalctl -u fluent-bit -f
+```
+
+## 系统要求
+
+- **操作系统**: Ubuntu/Debian/CentOS/RHEL/Fedora
+- **架构**: x86_64/amd64
+- **权限**: root 权限（用于安装和卸载）
+- **依赖**: curl（用于健康检查）
+
+## 版本信息
+
+- **Fluent Bit 版本**: 3.1.9
+- **安装包版本**: 1.0
+- **支持架构**: amd64
+
+## 注意事项
+
+1. 安装前请确保系统已更新
+2. 卸载时会保留 fluent-bit 用户（系统用户，可能被其他服务使用）
+3. 配置文件包含环境变量，请根据实际环境调整
+4. 建议在生产环境使用前进行充分测试

src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/check_health.sh

@@ -0,0 +1,69 @@
+#!/bin/bash
+
+# Fluent Bit 健康检查脚本
+# 输出 JSON 格式结果
+
+set -e
+
+# 检查 Fluent Bit 健康状态
+check_health() {
+    local name="fluent-bit"
+    local status="unhealth"
+    local reason=""
+    local install_record="/opt/argus-metric/current/.install_record"
+    
+    # 首先尝试通过安装记录文件检查进程
+    if [[ -f "$install_record" ]]; then
+        # 尝试使用jq解析JSON格式的安装记录文件
+        local pid=""
+        if command -v jq &> /dev/null; then
+            pid=$(jq -r '.components."fluent-bit".pid // empty' "$install_record" 2>/dev/null || echo "")
+        else
+            # 如果没有jq，使用简单的文本解析方法
+            pid=$(grep -A 10 '"fluent-bit"' "$install_record" | grep '"pid"' | cut -d'"' -f4 | head -1)
+        fi
+        
+        if [[ -n "$pid" && "$pid" =~ ^[0-9]+$ ]]; then
+            if kill -0 "$pid" 2>/dev/null; then
+                # 进程存在且运行正常
+                status="health"
+                reason="进程运行正常 (PID: $pid)"
+                echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+                exit 0
+            else
+                reason="安装记录中的 PID $pid 进程不存在"
+                echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+                exit 1
+            fi
+        else
+            reason="安装记录文件中未找到有效的 fluent-bit PID"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    else
+        # 如果安装记录文件不存在，尝试查找 fluent-bit 进程
+        local pids=$(pgrep -f "fluent-bit" 2>/dev/null || true)
+        if [[ -n "$pids" ]]; then
+            # 取第一个找到的 PID
+            local pid=$(echo "$pids" | head -1)
+            status="health"
+            reason="发现 fluent-bit 进程运行 (PID: $pid)，但未找到安装记录"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 0
+        else
+            reason="未找到 fluent-bit 进程，且安装记录文件不存在"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    fi
+}
+
+# 主函数
+main() {
+    check_health
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi

src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/fluent-bit.conf

@@ -0,0 +1,37 @@
+[SERVICE]
+    Daemon       Off
+    Parsers_File parsers.conf
+    HTTP_Server  On
+    HTTP_Listen  0.0.0.0
+    HTTP_Port    2020
+    storage.path /buffers
+    storage.sync normal
+    storage.checksum on
+    storage.backlog.mem_limit 128M
+    # 备注：该镜像默认未开启 Hot Reload，修改配置后请重启容器。
+
+@INCLUDE inputs.d/*.conf
+
+[FILTER]
+    Name   parser
+    Match  app.*
+    Key_Name log
+    Parser timestamp_parser
+    Reserve_Data On
+    Preserve_Key On
+    Unescape_Key On
+
+[FILTER]
+    Name   record_modifier
+    Match  *
+    Record cluster  ${CLUSTER}
+    Record rack     ${RACK}
+    Record host     ${HOSTNAME}
+
+[FILTER]
+    Name   lua
+    Match  app.*
+    script inject_labels.lua
+    call   add_labels
+
+@INCLUDE outputs.d/*.conf

src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inject_labels.lua

@@ -0,0 +1,15 @@
+function add_labels(tag, ts, record)
+  record["job_id"] = os.getenv("FB_JOB_ID") or record["job_id"] or "unknown"
+  record["user"]   = os.getenv("FB_USER")   or record["user"]   or "unknown"
+  record["model"]  = os.getenv("FB_MODEL")  or record["model"]  or "unknown"
+  record["gpu_id"] = os.getenv("FB_GPU_ID") or record["gpu_id"] or "na"
+  local p = record["log_path"] or ""
+  if string.find(p, "/logs/infer/") then
+    record["role"] = "infer"
+  elseif string.find(p, "/logs/train/") then
+    record["role"] = "train"
+  else
+    record["role"] = record["role"] or "app"
+  end
+  return 1, ts, record
+end