[#37 ] 测试通过swarm跨机部署节点；更新文档

[#37 ] server install 增加重试自检
[#37 ] swarm 部署优化
2025-11-05 09:57:08 +08:00 · 2025-11-05 09:57:08 +08:00 · 2025-11-05 09:57:08 +08:00 · 2025-11-05 09:57:08 +08:00 · 2025-11-05 09:57:08 +08:00 · 2025-11-05 09:57:07 +08:00
208 changed files with 13396 additions and 1207 deletions
--- a/.gitattributes
+++ b/.gitattributes
@@ -0,0 +1 @@
+src/metric/client-plugins/all-in-one-full/plugins/*/bin/* filter=lfs diff=lfs merge=lfs -text
--- a/build/build_images.sh
+++ b/build/build_images.sh
@@ -10,20 +10,28 @@ 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
@@ -41,10 +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
@@ -115,14 +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
@@ -131,29 +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 ""
@@ -184,6 +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 "======================================="
@@ -210,7 +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"
--- a/build/save_images.sh
+++ b/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
--- a/deployment/.gitignore
+++ b/deployment/.gitignore
@@ -0,0 +1 @@
+artifact/
--- a/deployment/build/README.md
+++ b/deployment/build/README.md
@@ -0,0 +1,16 @@
+# Deployment Build Toolkit
+
+This folder provides scripts to produce offline server/client packages and publish the client package to FTP.
+
+Commands
+- build_server_package.sh [--version YYYYMMDD]
+- build_client_package.sh [--version YYYYMMDD]
+- publish_client.sh --version YYYYMMDD --server <host> --user ftpuser --password <pass> [--port 21]
+
+Outputs
+- deployment/artifact/server/<YYYYMMDD>/
+- deployment/artifact/client/<YYYYMMDD>/
+
+Notes
+- Server package contains docker images (single all-images.tar.gz), compose/, scripts/, docs/, private/ skeleton.
+- Client package reuses all-in-one-full artifact, repacked as argus-metric_<YYYYMMDD>.tar.gz (compatible with setup.sh).
--- a/deployment/build/build_client_package.sh
+++ b/deployment/build/build_client_package.sh
@@ -0,0 +1,90 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
+BUILD_DIR="$ROOT_DIR/deployment/build"
+ART_ROOT="$ROOT_DIR/deployment/artifact"
+
+. "$BUILD_DIR/common.sh"
+
+usage() { cat <<'EOF'
+Build Argus Client Offline Package
+
+Usage: build_client_package.sh [--version YYYYMMDD] [--out DIR]
+
+Produces: deployment/artifact/client/<YYYYMMDD>/argus-metric_<YYYYMMDD>.tar.gz
+EOF
+}
+
+VERSION="$(today_version)"
+OUT_DIR=""
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --version) VERSION="$2"; shift 2;;
+    --out) OUT_DIR="$2"; shift 2;;
+    -h|--help) usage; exit 0;;
+    *) err "unknown arg: $1"; usage; exit 1;;
+  esac
+done
+
+PKG_DIR="${OUT_DIR:-$ART_ROOT/client/$VERSION}"
+make_dir "$PKG_DIR"
+
+log "Packaging client from all-in-one-full artifact"
+PLUGIN_DIR="$ROOT_DIR/src/metric/client-plugins/all-in-one-full"
+require_cmd bash tar gzip
+
+(cd "$PLUGIN_DIR" && bash scripts/package_artifact.sh --force)
+
+# pick latest artifact dir
+ART_BASE="$PLUGIN_DIR/artifact"
+latest_dir=$(ls -1dt "$ART_BASE"/*/ 2>/dev/null | head -n1 || true)
+[[ -n "$latest_dir" ]] || { err "no client artifact found in $ART_BASE"; exit 1; }
+
+tmpdir=$(mktemp -d)
+trap 'rm -rf "$tmpdir"' EXIT
+# Filter-only copy: keep install_order files + scripts + deps + version.json
+mkdir -p "$tmpdir/src"
+cp -f "$latest_dir/version.json" "$tmpdir/src/version.json"
+if command -v jq >/dev/null 2>&1; then
+  mapfile -t files < <(jq -r '.install_order[]' "$latest_dir/version.json")
+else
+  files=( $(grep -E '"install_order"' -A10 "$latest_dir/version.json" | sed -n 's/\s*"\(.*\.tar\.gz\)".*/\1/p') )
+fi
+for f in "${files[@]}"; do
+  [[ -f "$latest_dir/$f" ]] && cp -f "$latest_dir/$f" "$tmpdir/src/$f"
+done
+for aux in install.sh uninstall.sh check_health.sh check_version.sh sync_dns.sh restart_unhealthy.sh config.env; do
+  [[ -f "$latest_dir/$aux" ]] && cp -f "$latest_dir/$aux" "$tmpdir/src/$aux";
+done
+if [[ -d "$latest_dir/deps" ]]; then
+  mkdir -p "$tmpdir/src/deps" && rsync -a "$latest_dir/deps/" "$tmpdir/src/deps/" >/dev/null 2>&1 || cp -r "$latest_dir/deps/." "$tmpdir/src/deps/";
+fi
+
+out_name="argus-metric_$(echo "$VERSION" | sed 's/\./_/g').tar.gz"
+
+(cd "$tmpdir/src" && tar -czf "$PKG_DIR/$out_name" .)
+
+log "Client package ready: $PKG_DIR/$out_name"
+echo "$VERSION" > "$PKG_DIR/LATEST_VERSION"
+
+# include publish helper and setup.sh for convenience (place first)
+PUBLISH_TPL="$BUILD_DIR/templates/client/publish.sh"
+if [[ -f "$PUBLISH_TPL" ]]; then
+  cp "$PUBLISH_TPL" "$PKG_DIR/publish.sh"
+fi
+
+# also place a copy of setup.sh alongside
+SETUP_SRC="$ROOT_DIR/src/metric/client-plugins/all-in-one-full/scripts/setup.sh"
+[[ -f "$SETUP_SRC" ]] && cp "$SETUP_SRC" "$PKG_DIR/setup.sh" || true
+
+# docs for end users (this may overwrite file modes), then fix execute bits
+CLIENT_DOC_DIR="$BUILD_DIR/templates/client"
+if [[ -d "$CLIENT_DOC_DIR" ]]; then
+  rsync -a "$CLIENT_DOC_DIR/" "$PKG_DIR/" >/dev/null 2>&1 || cp -r "$CLIENT_DOC_DIR/." "$PKG_DIR/"
+fi
+
+# ensure helpers are executable
+chmod +x "$PKG_DIR/publish.sh" "$PKG_DIR/setup.sh" 2>/dev/null || true
+
+exit 0
--- a/deployment/build/build_server_package.sh
+++ b/deployment/build/build_server_package.sh
@@ -0,0 +1,138 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
+BUILD_DIR="$ROOT_DIR/deployment/build"
+ART_ROOT="$ROOT_DIR/deployment/artifact"
+
+. "$BUILD_DIR/common.sh"
+
+usage() { cat <<'EOF'
+Build Argus Server Offline Package
+
+Usage: build_server_package.sh [--version YYYYMMDD] [--out DIR] [--resave-image]
+
+Outputs into deployment/artifact/server/<YYYYMMDD>/ by default.
+EOF
+}
+
+VERSION="$(today_version)"
+OUT_DIR=""
+RESAVE_IMAGE=false
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --version) VERSION="$2"; shift 2;;
+    --out) OUT_DIR="$2"; shift 2;;
+    --resave-image) RESAVE_IMAGE=true; shift;;
+    -h|--help) usage; exit 0;;
+    *) err "unknown arg: $1"; usage; exit 1;;
+  esac
+done
+
+PKG_DIR="${OUT_DIR:-$ART_ROOT/server/$VERSION}"
+STAGE="$(mktemp -d)"
+trap 'rm -rf "$STAGE"' EXIT
+
+log "Version: $VERSION"
+log "Staging: $STAGE"
+
+# 1) Layout
+make_dir "$STAGE/images"
+make_dir "$STAGE/compose"
+make_dir "$STAGE/scripts"
+make_dir "$STAGE/docs"
+make_dir "$STAGE/private/argus"
+
+# 2) Compose: derive from sys/tests by removing test-only services
+SRC_COMPOSE="$ROOT_DIR/src/sys/tests/docker-compose.yml"
+[[ -f "$SRC_COMPOSE" ]] || { err "missing $SRC_COMPOSE"; exit 1; }
+# 2.1 filter out test services
+tmp_compose1="$STAGE/compose/docker-compose.filtered.yml"
+awk -f "$BUILD_DIR/templates/docker-compose.filter.awk" -v remove="node-a,node-b,test-node,test-gpu-node" "$SRC_COMPOSE" > "$tmp_compose1"
+# 2.2 transform to external overlay network (remove sysnet and per-service blocks)
+awk -f "$BUILD_DIR/templates/docker-compose.overlay.awk" "$tmp_compose1" > "$STAGE/compose/docker-compose.yml"
+cp "$BUILD_DIR/templates/.env.example" "$STAGE/compose/.env.example"
+# fix relative private path to match package layout (compose/ and private/ are siblings)
+sed -i 's#\./private/#../private/#g' "$STAGE/compose/docker-compose.yml"
+# also handle bind mount form without trailing slash
+sed -i -E 's#- \./private:/private#- ../private:/private#g' "$STAGE/compose/docker-compose.yml"
+# drop timezone file bind which may not exist on target distros (e.g. NixOS)
+sed -i '/\/etc\/timezone:\/etc\/timezone:ro/d' "$STAGE/compose/docker-compose.yml"
+
+# sanity-check: ensure test services are absent and external network present
+if grep -E '^(\s{2})(node-a|node-b|test-node|test-gpu-node):\s*$' -n "$STAGE/compose/docker-compose.yml" >/dev/null; then
+  err "compose filter failed: test services still present"; exit 1;
+fi
+if ! grep -q '^networks:' "$STAGE/compose/docker-compose.yml" || ! grep -q 'argus-sys-net:' "$STAGE/compose/docker-compose.yml"; then
+  err "compose overlay transform failed: external network missing"; exit 1;
+fi
+
+# 3) Images (reuse if already exported unless --resave-image)
+existing_images_tar="$PKG_DIR/images/all-images.tar.gz"
+if [[ "$RESAVE_IMAGE" == false && -f "$existing_images_tar" ]]; then
+  log "Reusing existing images tar: $existing_images_tar"
+  cp "$existing_images_tar" "$STAGE/images/"
+elif [[ "$RESAVE_IMAGE" == false ]]; then
+  # Try cross-version reuse from latest server_*.tar.gz
+  latest_pkg=$(ls -1t "$ART_ROOT/server"/server_*.tar.gz 2>/dev/null | head -n1 || true)
+  if [[ -n "$latest_pkg" ]]; then
+    log "Reusing images from: $latest_pkg"
+    mkdir -p "$STAGE/images"
+    # extract matching file regardless of top-level dir
+    if tar -xzf "$latest_pkg" -C "$STAGE" --wildcards '*/images/all-images.tar.gz' 2>/dev/null; then
+      # locate and move
+      found=$(find "$STAGE" -type f -path '*/images/all-images.tar.gz' | head -n1 || true)
+      if [[ -n "$found" ]]; then
+        mv "$found" "$STAGE/images/all-images.tar.gz"
+        # cleanup leftover extracted dir
+        dir_to_clean=$(dirname "$found")
+        rm -rf "${dir_to_clean%/images}" 2>/dev/null || true
+      fi
+    fi
+  fi
+fi
+
+# If still not present, save from local docker daemon
+if [[ ! -f "$STAGE/images/all-images.tar.gz" ]]; then
+  require_cmd docker gzip
+  images=(
+    argus-bind9:latest
+    argus-master:latest
+    argus-elasticsearch:latest
+    argus-kibana:latest
+    argus-metric-ftp:latest
+    argus-metric-prometheus:latest
+    argus-metric-grafana:latest
+    argus-alertmanager:latest
+    argus-web-frontend:latest
+    argus-web-proxy:latest
+  )
+  log "Saving images: ${#images[@]}"
+  tarfile="$STAGE/images/all-images.tar"
+  docker save -o "$tarfile" "${images[@]}"
+  gzip -f "$tarfile"
+fi
+
+# 4) Scripts & Docs
+copy_tree "$BUILD_DIR/templates/scripts" "$STAGE/scripts"
+copy_tree "$BUILD_DIR/templates/docs" "$STAGE/docs"
+
+# 5) Manifests
+gen_manifest "$STAGE" "$STAGE/manifest.txt"
+checksum_dir "$STAGE" "$STAGE/checksums.txt"
+
+# 6) Move to artifact
+make_dir "$PKG_DIR"
+rsync -a "$STAGE/" "$PKG_DIR/" 2>/dev/null || cp -r "$STAGE/." "$PKG_DIR/"
+log "Server package ready: $PKG_DIR"
+
+echo "$VERSION" > "$PKG_DIR/version.json"
+
+# 7) Create distributable tarball
+OUT_TAR_DIR="$(dirname "$PKG_DIR")"
+OUT_TAR="$OUT_TAR_DIR/server_${VERSION}.tar.gz"
+log "Creating tarball: $OUT_TAR"
+(cd "$PKG_DIR/.." && tar -czf "$OUT_TAR" "$(basename "$PKG_DIR")")
+log "Tarball ready: $OUT_TAR"
+
+exit 0
--- a/deployment/build/common.sh
+++ b/deployment/build/common.sh
@@ -0,0 +1,33 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+log() { echo -e "\033[0;34m[INFO]\033[0m $*"; }
+warn() { echo -e "\033[1;33m[WARN]\033[0m $*"; }
+err()  { echo -e "\033[0;31m[ERR ]\033[0m $*" >&2; }
+
+require_cmd() {
+  for c in "$@"; do
+    command -v "$c" >/dev/null 2>&1 || { err "missing command: $c"; exit 1; }
+  done
+}
+
+today_version() {
+  date +%Y%m%d
+}
+
+checksum_dir() {
+  local dir="$1"; local out="$2"; : > "$out";
+  (cd "$dir" && find . -type f -print0 | sort -z | xargs -0 sha256sum) >> "$out"
+}
+
+make_dir() { mkdir -p "$1"; }
+
+copy_tree() {
+  local src="$1" dst="$2"; rsync -a --delete "$src/" "$dst/" 2>/dev/null || cp -r "$src/." "$dst/";
+}
+
+gen_manifest() {
+  local root="$1"; local out="$2"; : > "$out";
+  (cd "$root" && find . -maxdepth 3 -type f -printf "%p\n" | sort) >> "$out"
+}
+
--- a/deployment/build/templates/.env.example
+++ b/deployment/build/templates/.env.example
@@ -0,0 +1,32 @@
+# UID/GID for service processes
+ARGUS_BUILD_UID=1000
+ARGUS_BUILD_GID=1000
+
+# Host ports (adjust if occupied)
+MASTER_PORT=32300
+ES_HTTP_PORT=9200
+KIBANA_PORT=5601
+NODE_A_PORT=2020
+NODE_B_PORT=2021
+PROMETHEUS_PORT=9090
+GRAFANA_PORT=3000
+ALERTMANAGER_PORT=9093
+WEB_PROXY_PORT_8080=8080
+WEB_PROXY_PORT_8081=8081
+WEB_PROXY_PORT_8082=8082
+WEB_PROXY_PORT_8083=8083
+WEB_PROXY_PORT_8084=8084
+WEB_PROXY_PORT_8085=8085
+
+# FTP
+FTP_PORT=21
+FTP_DATA_PORT=20
+FTP_PASSIVE_HOST_RANGE=21100-21110
+FTP_PASSWORD=ZGClab1234!
+FTP_DOMAIN=ftp.metric.argus.com
+
+# GPU profile disabled by default
+ENABLE_GPU=false
+
+# External overlay network (Swarm attachable)
+OVERLAY_NET_NAME=argus-sys-net
--- a/deployment/build/templates/client/INSTALL_CLIENT_zh.md
+++ b/deployment/build/templates/client/INSTALL_CLIENT_zh.md
@@ -0,0 +1,44 @@
+# Argus Metric 客户端安装指南（容器内普通用户场景）
+
+## 准备与连通性检查
+- FTP 连接（需要账号密码，默认 `ftpuser / ZGClab1234!`）
+  - `curl -u ftpuser:ZGClab1234! -I ftp://<FTP_IP>:21/LATEST_VERSION`
+  - `curl -u ftpuser:ZGClab1234! -s ftp://<FTP_IP>:21/ | head`
+- 下载安装脚本
+  - `curl -u ftpuser:ZGClab1234! -fsSL ftp://<FTP_IP>:21/setup.sh -o /tmp/setup.sh`
+  - `chmod +x /tmp/setup.sh`
+
+## 元数据与主机名
+- Agent 需要元数据（env/user/instance）与 Master 地址：
+  - 方式A：hostname 形如 `env-user-instance-xxx`（推荐）
+  - 方式B：导出环境变量：
+    - `export AGENT_ENV=dev`
+    - `export AGENT_USER=<your_user>`
+    - `export AGENT_INSTANCE=<node_id>`
+- Master 地址：
+  - `export MASTER_ENDPOINT=http://master.argus.com:3000`
+
+> 安装脚本会在开头检查上述条件，缺失时会终止并给出修复提示。
+
+## 执行安装
+- 以 root 运行（容器内如为非 root 用户请切换为 root）：
+  - `sudo /tmp/setup.sh --server <FTP_IP> --user ftpuser --password 'ZGClab1234!' --port 21`
+- 如需自定义安装根目录：`--install-dir /opt/argus-metric`
+
+提示（容器接入 overlay 网络时）：
+- 在执行 setup 前，先将容器内 DNS 指向 Bind9 的 overlay IP：
+  - `echo "nameserver <BIND_OVERLAY_IP>" > /etc/resolv.conf`
+  - 这样 `master.argus.com`、`es.log.argus.com` 等域名即可解析；首次下载 `setup.sh` 仍建议使用 FTP 的 overlay IP。
+
+更多快速步骤请参考：`QUICK_NODE_DEPLOY_zh.md`。
+
+## 安装后自检（setup 自动执行）
+- setup 会等待最多 5 分钟，确认以下条件后才报告完成：
+  - `/private/argus/agent/<hostname>/node.json` 已生成；
+  - `last_report` 在持续更新；
+  - `health.metric-argus-agent|metric-node-exporter|metric-fluent-bit|metric-dcgm-exporter` 均为 `healthy` 且 `error` 为空。
+
+## 手工验证（可选）
+- `cat /private/argus/agent/$(hostname)/node.json | jq '.'`
+- `curl -s -o /dev/null -w "%{http_code}\n" http://localhost:9100/metrics` 应为 200
+- 查看日志：`/var/log/argus-agent.log`、`/opt/argus-metric/versions/*/.install.log`
--- a/deployment/build/templates/client/PUBLISH_CLIENT_zh.md
+++ b/deployment/build/templates/client/PUBLISH_CLIENT_zh.md
@@ -0,0 +1,57 @@
+# Argus Metric 客户端发布说明（FTP）
+
+本说明面向“发布人员”，讲清楚如何把客户端离线包发布到 FTP，供各节点通过 `curl` 自动安装。
+
+## 目录结构（构建后）
+- `client-YYYYMMDD/`
+  - `argus-metric_YYYYMMDD.tar.gz` 客户端离线包
+  - `setup.sh` 客户端安装入口脚本（提供给节点用 curl 下载）
+  - `publish.sh` 发布脚本（将上述两项与 `LATEST_VERSION` 上传到 FTP）
+  - `LATEST_VERSION` 文本（内容为 `YYYYMMDD`，或 `YYYYMMDD-rN`）
+  - `INSTALL_CLIENT_zh.md` 本地安装指南（给使用者看，不会上载到 FTP）
+  - `PUBLISH_CLIENT_zh.md` 本说明
+
+> 注意：`publish.sh`/`setup.sh` 为可执行脚本；构建脚本已保证二者具有执行权限。
+
+## 前置条件
+- FTP 服务已运行（默认容器：`argus-ftp`），并打开端口：21、20、21100–21110（被动模式）。
+- FTP 账号：默认 `ftpuser / ZGClab1234!`（如有更改，以实际为准）。
+
+## 发布步骤（在 server 机器或能直连 FTP 的任意机器上）
+1) 进入发布目录：
+- `cd client-YYYYMMDD`
+
+2) 执行发布：
+- `./publish.sh --server <FTP_HOST> --user <USER> --password '<PASS>' [--port 21]`
+- 例如在服务端本机：`./publish.sh --server localhost --user ftpuser --password 'ZGClab1234!' --port 21`
+
+脚本会上传三类文件到 FTP 根：
+- `setup.sh`
+- `argus-metric_YYYYMMDD[ -rN ].tar.gz`
+- `LATEST_VERSION`（内容为当前版本号）
+
+3) 发布后验证：
+- `curl -u ftpuser:****** -I ftp://<FTP_HOST>:21/LATEST_VERSION` 应返回 200
+- `curl -u ftpuser:****** -fsSL ftp://<FTP_HOST>:21/LATEST_VERSION` 内容为版本号（如 `20251104`）
+- `curl -u ftpuser:****** -I ftp://<FTP_HOST>:21/argus-metric_YYYYMMDD.tar.gz` 返回 200
+
+## 节点侧使用方式（摘要）
+- 首次下载用 FTP 的“IP 地址”：
+  - `curl -u ftpuser:****** -fsSL ftp://<FTP_IP>:21/setup.sh -o /tmp/setup.sh && chmod +x /tmp/setup.sh`
+- 执行安装：
+  - 必需元数据：`AGENT_ENV/AGENT_USER/AGENT_INSTANCE`，以及 `MASTER_ENDPOINT=http://master.argus.com:3000`
+  - `sudo /tmp/setup.sh --server <FTP_IP> --user ftpuser --password '******' --port 21`
+- overlay 容器场景：
+  - 先将容器内 DNS 指向 Bind9 的 overlay IP：`echo "nameserver <BIND_OVERLAY_IP>" > /etc/resolv.conf`
+  - 然后再执行上述安装；安装后约 1–2 分钟内 DNS 即可解析 `*.argus.com` 域名。
+
+## 常见问题
+- `530 Access denied`：用户名/密码错误或 FTP 目录无权限；请核对账号与 FTP 容器状态。
+- `Permission denied` 执行 `publish.sh`：为脚本权限问题；`chmod +x publish.sh`。构建脚本已修复默认权限。
+- 被动端口不通导致失败：请开放 21100–21110。
+- 客户端安装后短时 `curl http://master.argus.com:3000` 为 000：服务冷启动或 DNS 同步延迟，等待 1–2 分钟再试。
+
+## 版本与回滚
+- `LATEST_VERSION` 决定客户端默认安装的版本号。
+- 如需回滚：将旧版本号写回 `LATEST_VERSION` 并重新发布（或手动指定 `--version` 安装）。
+
--- a/deployment/build/templates/client/QUICK_NODE_DEPLOY_zh.md
+++ b/deployment/build/templates/client/QUICK_NODE_DEPLOY_zh.md
@@ -0,0 +1,58 @@
+# Argus Metric 节点快速部署（Overlay 网络容器）
+
+本文档给出在 Docker Swarm external overlay 网络中，快速拉起一个测试节点并完成注册的最小可行步骤。
+
+## 前提
+- 服务端已在 Manager 机安装完成并运行良好（`server-selfcheck` 通过）。
+- Overlay 网络名称：`argus-sys-net`（默认）。
+- 已通过 FTP 发布 `setup.sh` 与客户端包，且能从 FTP 获取 `LATEST_VERSION`。
+- 用于测试的镜像：`argus-sys-metric-test-node:latest` 已存在于目标机器。
+
+## 步骤
+
+- 获取 FTP 和 Bind 的 overlay IP（在 Manager 上执行）
+  - `FTPIP=$(docker inspect -f '{{ (index .NetworkSettings.Networks "argus-sys-net").IPAddress }}' argus-ftp)`
+  - `BINDIP=$(docker inspect -f '{{ (index .NetworkSettings.Networks "argus-sys-net").IPAddress }}' argus-bind-sys)`
+  - `echo "FTP=$FTPIP BIND=$BINDIP"`
+
+- 准备宿主挂载目录（以 s4 为例）
+  - `mkdir -p /home2/yuyr/deploy/test-metric-node/s4`
+
+- 启动测试节点容器（接入 overlay）
+  - `docker run -d --name argus-metric-test-node-s4 \
+      --hostname dev2-yuyr-node002s4 \
+      --network argus-sys-net \
+      -v /home2/yuyr/deploy/test-metric-node/s4:/private/argus/agent \
+      argus-sys-metric-test-node:latest sleep infinity`
+
+- 在容器内执行安装（先用 FTP IP 引导，DNS 指向 Bind）
+  - `docker exec -it argus-metric-test-node-s4 bash`
+  - `echo "nameserver $BINDIP" > /etc/resolv.conf`
+  - `curl --ftp-method nocwd -u ftpuser:ZGClab1234! -fsSL "ftp://$FTPIP:21/setup.sh" -o /tmp/setup.sh`
+  - `chmod +x /tmp/setup.sh`
+  - `export AGENT_ENV=dev2 AGENT_USER=yuyr AGENT_INSTANCE=node002s4`
+  - `export MASTER_ENDPOINT=http://master.argus.com:3000`
+  - `/tmp/setup.sh --server "$FTPIP" --user ftpuser --password 'ZGClab1234!' --port 21`
+  - 说明：setup 会自动执行安装后自检（最多 5 分钟），无需手动轮询。
+
+## 验证（推荐在容器内执行，避免宿主权限问题）
+
+- 查看 node.json 关键字段
+  - `cat /private/argus/agent/dev2-yuyr-node002s4/node.json | jq '{last_report, health}'`
+  - 期望：四个 health 全部 healthy；等待 ≥70s 再查看，`last_report` 持续更新。
+
+- 指标端口
+  - `curl -s -o /dev/null -w '%{http_code}\n' http://localhost:9100/metrics`（期望 200）
+  - （如测试 GPU）`curl -s -o /dev/null -w '%{http_code}\n' http://localhost:9400/metrics`（有 GPU 时 200）
+
+- 与服务端连通（域名经 Bind 解析）
+  - `curl -s -o /dev/null -w '%{http_code}\n' http://master.argus.com:3000/readyz`（期望 200）
+  - `curl -s -o /dev/null -w '%{http_code}\n' http://es.log.argus.com:9200/_cluster/health`（期望 200）
+
+## （可选）在服务器主机侧观察 Prometheus 目标更新
+- `cat /home2/yuyr/deploy/versions/<VERSION>/private/argus/metric/prometheus/nodes.json | jq '.'`
+
+## 常见提示
+- 初次安装后短时 `curl` 域名返回 000/超时属正常，多等待 1–2 分钟 DNS 同步/组件冷启动完成。
+- 如在宿主直接读取挂载的 node.json 报 Permission denied，请使用 `docker exec` 在容器内查看。
+- MASTER_ENDPOINT 固定使用域名 `http://master.argus.com:3000`，客户端无需固定 IP。
--- a/deployment/build/templates/client/publish.sh
+++ b/deployment/build/templates/client/publish.sh
@@ -0,0 +1,54 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+usage() { cat <<'EOF'
+Publish Argus client package to FTP
+
+Usage:
+  ./publish.sh --server HOST --user USER --password PASS [--port 21]
+
+Notes:
+- This script expects to run inside the built client artifact directory.
+- It reads LATEST_VERSION and uploads setup.sh, argus-metric_<ver>.tar.gz, and LATEST_VERSION.
+EOF
+}
+
+HOST=""; USERNAME=""; PASSWORD=""; PORT=21
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --server) HOST="$2"; shift 2;;
+    --user) USERNAME="$2"; shift 2;;
+    --password) PASSWORD="$2"; shift 2;;
+    --port) PORT="$2"; shift 2;;
+    -h|--help) usage; exit 0;;
+    *) echo "unknown arg: $1" >&2; usage; exit 1;;
+  esac
+done
+
+[[ -n "$HOST" && -n "$USERNAME" && -n "$PASSWORD" ]] || { usage; exit 1; }
+
+here="$(pwd)"
+if [[ ! -f "$here/LATEST_VERSION" ]]; then
+  echo "LATEST_VERSION not found in $(pwd)" >&2; exit 1;
+fi
+VER=$(cat "$here/LATEST_VERSION" | tr -d '\n')
+PKG="argus-metric_${VER}.tar.gz"
+
+if [[ ! -f "$here/$PKG" ]]; then
+  echo "client tar not found: $PKG" >&2; exit 1
+fi
+
+# locate setup.sh (prefer colocated, fallback to bundled path if provided)
+SETUP="${here}/setup.sh"
+if [[ ! -f "$SETUP" ]]; then
+  echo "setup.sh not found in $(pwd)" >&2; exit 1
+fi
+
+echo "[PUBLISH] server=$HOST port=$PORT version=$VER"
+
+curl -u "$USERNAME:$PASSWORD" -sfT "$SETUP" "ftp://$HOST:$PORT/setup.sh"
+curl -u "$USERNAME:$PASSWORD" -sfT "$PKG"   "ftp://$HOST:$PORT/$PKG"
+printf "%s" "$VER" | curl -u "$USERNAME:$PASSWORD" -sfT - "ftp://$HOST:$PORT/LATEST_VERSION"
+
+echo "[OK] publish completed"
+
--- a/deployment/build/templates/docker-compose.filter.awk
+++ b/deployment/build/templates/docker-compose.filter.awk
@@ -0,0 +1,41 @@
+#!/usr/bin/awk -f
+# Remove specific service blocks from a docker-compose.yml by service name.
+# Usage: awk -f docker-compose.filter.awk -v remove="node-a,node-b,test-node,test-gpu-node" input.yml > output.yml
+
+BEGIN{
+  split(remove, rm, ",");
+  for(i in rm){
+    gsub(/^\s+|\s+$/,"",rm[i]);
+    if (rm[i] != "") skipname[rm[i]] = 1;
+  }
+  in_services=0; skipping=0;
+}
+
+function service_header(line,   m) {
+  # match exactly two leading spaces followed by name:
+  if (match(line, /^  ([A-Za-z0-9_-]+):[ ]*$/, m)) return m[1];
+  return "";
+}
+
+{
+  # Track top-level sections (no indentation)
+  if ($0 ~ /^[A-Za-z0-9_-]+:[ ]*$/ && $0 !~ /^\s/) {
+    in_services = ($0 ~ /^services:[ ]*$/) ? 1 : 0;
+  }
+
+  if (skipping) {
+    # Stop skipping at next service header or another top-level section
+    if (service_header($0) != "" || ($0 ~ /^(networks|volumes):[ ]*$/ && $0 !~ /^\s/)) {
+      skipping=0;
+    } else {
+      next;
+    }
+  }
+
+  if (in_services) {
+    name = service_header($0);
+    if (name != "" && (name in skipname)) { skipping=1; next; }
+  }
+
+  print;
+}
--- a/deployment/build/templates/docker-compose.overlay.awk
+++ b/deployment/build/templates/docker-compose.overlay.awk
@@ -0,0 +1,74 @@
+#!/usr/bin/awk -f
+# Transform docker-compose.yml to use an external overlay network for all services
+# - Remove top-level networks definition
+# - Remove per-service networks block (including ipv4_address and sysnet refs)
+# - Insert per-service networks: [argus-sys-net]
+# - Append external networks mapping at the end
+
+BEGIN{
+  in_top_networks=0; in_services=0; in_service=0; svc_indent=0; curr_name="";
+}
+
+function is_service_header(line){ return svc_name(line)!=""; }
+function svc_name(line,   m){ if (match(line, /^  ([A-Za-z0-9_-]+):[ ]*$/, m)) return m[1]; return ""; }
+
+function indent_len(s,   n){ n=match(s,/[^ ]/)-1; if(n<0) n=0; return n; }
+
+{
+  # Detect entry into top-level sections
+  if ($0 ~ /^[A-Za-z0-9_-]+:[ ]*$/ && $0 !~ /^\s/) {
+    in_services = ($0 ~ /^services:[ ]*$/);
+    # If a new top-level section starts, stop skipping top networks
+    in_top_networks = 0;
+  }
+
+  # Handle removal of initial top-level 'networks:' block
+  if ($0 ~ /^networks:[ ]*$/ && $0 !~ /^\s/) {
+    in_top_networks = 1; next;
+  }
+  if (in_top_networks) {
+    # skip until next top-level section (non-indented key)
+    next;
+  }
+
+  if (in_services) {
+    # Track service boundaries
+    if (is_service_header($0)) {
+      in_service=1; svc_indent=2; networks_inserted=0; curr_name=svc_name($0); print; next;
+    }
+    if (in_service) {
+      # If line is indented <= service indent, we've left this service
+      if (indent_len($0) <= svc_indent && $0 !~ /^\s*$/) {
+        in_service=0;
+      }
+    }
+
+    if (in_service) {
+      # Skip any existing networks block under the service
+      if ($0 ~ /^\s{4}networks:[ ]*$/) { skipping_nets=1; next; }
+      if (skipping_nets) {
+        if (indent_len($0) <= 4) { skipping_nets=0; }
+        else next;
+      }
+
+      # After container_name or image, inject networks once
+      if (!networks_inserted && ($0 ~ /^\s{4}container_name:/ || $0 ~ /^\s{4}image:/)) {
+        print; 
+        print "    networks:";
+        print "      - argus-sys-net";
+        networks_inserted=1; next;
+      }
+      # no host port injection; bind serves DNS inside overlay only
+    }
+  }
+
+  print;
+}
+
+END{
+  print "";
+  print "networks:";
+  print "  argus-sys-net:";
+  print "    external: true";
+  print "    name: ${OVERLAY_NET_NAME:-argus-sys-net}";
+}
--- a/deployment/build/templates/docs/INSTALL_SERVER.md
+++ b/deployment/build/templates/docs/INSTALL_SERVER.md
@@ -0,0 +1,50 @@
+# Argus Server Offline Installation
+
+## Prerequisites
+- Linux x86_64 (Ubuntu 22.04 recommended; see OS compatibility for NixOS)
+- Docker & Docker Compose installed
+- Open ports: 32300, 9200, 5601, 9090, 9093, 8080..8085, 21, 20, 21100–21110 (or auto-fallback to high ports)
+
+## Quick Start
+1. Extract to a target dir, e.g. `/opt/argus-deploy/versions/<YYYYMMDD>`
+2. `./server-install.sh` (non‑root is supported: it will precreate minimal dirs and auto-fix Kibana/ES/Bind in containers)
+3. `./server-status.sh`
+4. `./server-selfcheck.sh` (on failure it auto-runs diagnose and writes logs under `logs/`)
+5. `./server-uninstall.sh` to tear down
+
+## What the Installer Does
+- Loads local images (`images/all-images.tar.gz`)
+- Generates OS-compat override (`security_opt: ["label=disable"]`, `userns_mode: host`, bind `tmpfs:/run/named`)
+- Starts server-only services: bind/master/es/kibana/ftp/prometheus/grafana/alertmanager/web-frontend/web-proxy
+- DNS Bootstrap:
+  - Ensure `/private/argus/etc/dns.conf` exists (write `172.31.0.2` if missing);
+  - Run `/private/argus/etc/update-dns.sh` in dependent containers so `/etc/resolv.conf` points to bind;
+  - Wait for `*.argus.com` hint files, then reload bind;
+  - Restart web‑proxy to re-render nginx resolver from `dns.conf`;
+- Writes `logs/selfcheck.json` as final summary
+
+## OS Compatibility
+- NixOS / non-xattr FS: containers run with `security_opt: ["label=disable"]` and `userns_mode: host`.
+- If you cannot use sudo, the installer will:
+  - create minimal data dirs (incl. `private/argus/log/{elasticsearch,kibana}`) with permissive perms when possible;
+  - ensure inside containers: Kibana `data` → `/private/argus/log/kibana`, Elasticsearch `data` → `/private/argus/log/elasticsearch`, and Bind `rndc.key` is generated.
+  (Manual pre-creation scripts are no longer required.)
+
+## Files & Layout
+- `compose/` (docker-compose.yml, .env)
+- `private/` (data mounts)
+- `scripts/` (install/uninstall/status/selfcheck/diagnose)
+- `logs/` (selfcheck + diagnose outputs)
+
+## Troubleshooting (Quick)
+- Run `./server-selfcheck.sh` → see `logs/selfcheck.json`
+- Run `./server-diagnose.sh` → produces timestamped logs:
+  - `logs/diagnose_details_YYYYMMDD-HHMMSSZ.log`
+  - `logs/diagnose_error_YYYYMMDD-HHMMSSZ.log`
+  And updates `diagnose_details.log`/`diagnose_error.log` to the latest
+- Error lines are tagged `[service][source]`, e.g. `[kibana][http] /api/status=503`
+
+Common issues:
+- Kibana 503: wait cold start or fix DNS so `es.log.argus.com` resolves
+- web‑proxy 504: check nginx `resolver` includes `172.31.0.2 127.0.0.11`
+- EACCES/locks: ensure `sudo ./server-prepare-dirs.sh` ran and ownership matches UID:GID
--- a/deployment/build/templates/docs/INSTALL_SERVER_zh.md
+++ b/deployment/build/templates/docs/INSTALL_SERVER_zh.md
@@ -0,0 +1,29 @@
+# Argus 服务端离线安装指南
+
+## 先决条件
+- Linux x86_64（推荐 Ubuntu 22.04；NixOS 见“兼容说明”）
+- 已安装 Docker 与 Docker Compose
+- 端口：32300, 9200, 5601, 9090, 9093, 8080..8085, 21, 20, 21100–21110
+
+## 快速开始
+1. 解压到目标目录（例如 `/opt/argus-deploy/versions/<YYYYMMDD>`）
+2. 安装：`./server-install.sh`（支持普通用户：会自动创建最小目录并在容器内修复 Kibana/ES/Bind）
+3. 状态：`./server-status.sh`
+4. 自检：`./server-selfcheck.sh`（失败会自动采集诊断）
+5. 卸载：`./server-uninstall.sh`
+
+## 安装流程要点
+- 仅启动 10 个服务端组件（不包含测试节点）；
+- DNS Bootstrap：补齐首次部署 DNS 依赖（生成/确认 `dns.conf`、统一容器 resolv.conf、写入 `*.argus.com`、reload bind、重启 web‑proxy）；
+- 输出自检结果到 `logs/selfcheck.json`。
+
+## 兼容说明（NixOS 等）
+- 使用 `security_opt: ["label=disable"]` 与 `userns_mode: host`。
+- 非 root 场景：安装器会创建最小目录（含 `private/argus/log/{elasticsearch,kibana}`），并在容器内完成：
+  - Kibana 的 `data` 软链到 `/private/argus/log/kibana`
+  - Elasticsearch 的 `data` 软链到 `/private/argus/log/elasticsearch`
+  - Bind 生成 `/etc/bind/rndc.key`
+
+## 故障排查（见下文 Troubleshooting_zh）
+- `./server-selfcheck.sh` → `logs/selfcheck.json`
+- `./server-diagnose.sh` → `logs/diagnose_error_*.log` / `logs/diagnose_details_*.log`
--- a/deployment/build/templates/docs/SWARM_DEPLOY_zh.md
+++ b/deployment/build/templates/docs/SWARM_DEPLOY_zh.md
@@ -0,0 +1,50 @@
+# Argus 多机部署（Docker Swarm + External Overlay）
+
+- 前提：Docker ≥ 20.10；Manager/Worker 节点开放 2377/tcp、7946/tcp+udp、4789/udp。
+- DNS：Bind9 为统一权威，解析 *.argus.com 至部署机固定对外主机 IP。
+
+## 在部署机（Manager）
+- 初始化 Swarm：`docker swarm init --advertise-addr <manager_ip>`
+- 创建 overlay：`docker network create --driver overlay --attachable argus-sys-net`
+- 解压离线包后执行：
+  - `./server-install.sh`（会验证 Swarm 状态、确保 overlay 存在、写入 dns.conf）
+  - `./server-selfcheck.sh`（失败会自动触发诊断）
+
+## 在节点机（Worker 或非 Docker 主机）
+- Swarm Worker：执行 Manager 的 `docker swarm join ...`；
+- 运行客户端容器：
+  - `docker run -d --name argus-metric-node-001 --network argus-sys-net -v /data/argus/agent:/private/argus/agent argus-sys-metric-test-node:latest sleep infinity`
+- 进入容器安装（先 IP 引导，后域名）：
+  - `curl -u ftpuser:*** -fsSL ftp://<FTP_IP>:21/setup.sh -o /tmp/setup.sh && chmod +x /tmp/setup.sh`
+  - `AGENT_ENV=dev AGENT_USER=alice AGENT_INSTANCE=node001 MASTER_ENDPOINT=http://master.argus.com:3000 /tmp/setup.sh --server ftp.argus.com --user ftpuser --password '***' --port 21`
+
+## 关键点
+- 首次必须使用 FTP 的 IP 引导（下载 setup.sh 与 dns.conf）
+- MASTER_ENDPOINT 永远使用域名：`http://master.argus.com:3000`
+- docker compose 改为 external overlay；容器内不使用 Docker 服务名；web-proxy 与组件上游统一用域名
+
+## 找回/轮换 Swarm 加入令牌与解锁密钥
+
+在任意一个 Manager 节点上执行以下命令即可查看或轮换加入令牌（join token）：
+
+- 查看加入 Worker 的命令：
+  - `docker swarm join-token worker`
+- 只打印 Worker 的 token：
+  - `docker swarm join-token -q worker`
+- 查看加入 Manager 的命令：
+  - `docker swarm join-token manager`
+- 只打印 Manager 的 token：
+  - `docker swarm join-token -q manager`
+
+在待加入节点执行（示例，替换 Manager_IP）：
+- `docker swarm join --token <上面查到的token> <Manager_IP>:2377`
+
+轮换 token（怀疑泄露或需要更新时）：
+- 轮换 Worker：`docker swarm join-token --rotate worker`
+- 轮换 Manager：`docker swarm join-token --rotate manager`
+
+如果你指的是“解锁密钥”（autolock 的 unlock key），在 Manager 上：
+- 查看：`docker swarm unlock-key`
+- 轮换：`docker swarm unlock-key --rotate`
+
+提示：当看到 “This node is not a swarm manager.” 时，说明当前节点不是 Manager，需要到 Manager 节点执行，或在现有 Manager 上 `docker node promote <NODE-ID>` 将其提升为 Manager。
--- a/deployment/build/templates/docs/TROUBLESHOOTING.md
+++ b/deployment/build/templates/docs/TROUBLESHOOTING.md
@@ -0,0 +1,20 @@
+# Troubleshooting
+
+- Status: `scripts/server-status.sh`
+- Selfcheck: `scripts/server-selfcheck.sh`
+- Diagnose: `scripts/server-diagnose.sh`
+
+Outputs:
+- `logs/selfcheck.json`
+- `logs/diagnose_details_*.log` (full details)
+- `logs/diagnose_error_*.log` (tagged errors)
+
+Web‑Proxy:
+- 8083 expects 200/302/403; 8084/8085 must include CORS header
+- nginx resolver should be `172.31.0.2 127.0.0.11`
+
+Kibana/ES:
+- Verify `es.log.argus.com` resolves inside Kibana
+
+Permissions:
+- The installer auto-creates minimal dirs and applies container-side fixes (Kibana/ES/Bind). If you still see EACCES/lock errors, rerun `./server-install.sh` and review diagnose logs.
--- a/deployment/build/templates/docs/TROUBLESHOOTING_zh.md
+++ b/deployment/build/templates/docs/TROUBLESHOOTING_zh.md
@@ -0,0 +1,16 @@
+# 故障排查
+
+- 状态：`scripts/server-status.sh`
+- 自检：`scripts/server-selfcheck.sh`
+- 诊断：`scripts/server-diagnose.sh`
+
+输出：
+- `logs/selfcheck.json`
+- `logs/diagnose_error_*.log`（错误摘要）
+- `logs/diagnose_details_*.log`（详细信息）
+
+Web‑Proxy：8083=200/302/403；8084/8085 需包含 CORS
+Kibana：确认可解析 `es.log.argus.com`
+权限：
+- 非 root 安装时，安装器会创建最小目录并在容器内修复 Kibana/ES/Bind；
+- 如仍有 `EACCES`/锁文件报错，先重跑 `./server-install.sh`（会重复容器内修复），并查看诊断日志。
--- a/deployment/build/templates/scripts/server-diagnose.sh
+++ b/deployment/build/templates/scripts/server-diagnose.sh
@@ -0,0 +1,147 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+
+ENV_FILE="$ROOT/compose/.env"; [[ -f "$ENV_FILE" ]] && set -a && source "$ENV_FILE" && set +a
+
+ts="$(date -u +%Y%m%d-%H%M%SZ)"
+LOG_DIR="$ROOT/logs"
+mkdir -p "$LOG_DIR" || true
+# Fallback to /tmp when logs dir is not writable
+if ! ( : > "$LOG_DIR/.w" 2>/dev/null ); then
+  LOG_DIR="/tmp/argus-logs"
+  mkdir -p "$LOG_DIR" || true
+fi
+DETAILS="$LOG_DIR/diagnose_details_${ts}.log"
+ERRORS="$LOG_DIR/diagnose_error_${ts}.log"
+: > "$DETAILS"; : > "$ERRORS"
+
+logd() { echo "$(date '+%F %T') $*" >> "$DETAILS"; }
+append_err() { echo "$*" >> "$ERRORS"; }
+
+http_code() { curl -s -o /dev/null -w "%{http_code}" "$1" || echo 000; }
+http_body_head() { curl -s --max-time 3 "$1" 2>/dev/null | sed -n '1,5p' || true; }
+header_val() { curl -s -D - -o /dev/null "$@" | awk -F': ' 'BEGIN{IGNORECASE=1}$1=="Access-Control-Allow-Origin"{gsub("\r","",$2);print $2}'; }
+
+section() {
+  local name="$1"; logd "===== [$name] ====="; }
+
+svc() {
+  local svc_name="$1"; local cname="$2"; shift 2
+  section "$svc_name ($cname)"
+  logd "docker ps:"; docker ps -a --format '{{.Names}}\t{{.Status}}\t{{.Image}}' | awk -v n="$cname" '$1==n' >> "$DETAILS" || true
+  logd "docker inspect (state/restartcount):"; docker inspect -f '{{.State.Status}} rc={{.RestartCount}} started={{.State.StartedAt}}' "$cname" >> "$DETAILS" 2>&1 || true
+  logd "last 200 container logs:"; docker logs --tail 200 "$cname" >> "$DETAILS" 2>&1 || true
+
+  # extract error lines from container logs
+  docker logs --tail 200 "$cname" 2>&1 | \
+    grep -Ei '\b(error|failed|fail|exception|panic|fatal|critical|unhealthy|permission denied|forbidden|refused|traceback|错误|失败)\b' | \
+    sed "s/^/[${svc_name}][container] /" >> "$ERRORS" || true
+
+  # supervisor status and logs
+  if docker exec "$cname" sh -lc 'command -v supervisorctl >/dev/null 2>&1' >/dev/null 2>&1; then
+    logd "supervisorctl status:"; docker exec "$cname" sh -lc 'supervisorctl status' >> "$DETAILS" 2>&1 || true
+    # iterate supervisor logs and collect tails + errors per file
+    local files
+    files=$(docker exec "$cname" sh -lc 'ls /var/log/supervisor/*.log 2>/dev/null' || true)
+    for f in $files; do
+      logd "tail -n 80 $f:"; docker exec "$cname" sh -lc "tail -n 80 $f 2>/dev/null || true" >> "$DETAILS" 2>&1 || true
+      docker exec "$cname" sh -lc "tail -n 200 $f 2>/dev/null" 2>/dev/null | \
+        grep -Ei '\b(error|failed|fail|exception|panic|fatal|critical|unhealthy|permission denied|forbidden|refused|traceback|错误|失败)\b' | \
+        sed "s/^/[${svc_name}][supervisor:$(basename $f)] /" >> "$ERRORS" || true
+    done
+  fi
+}
+
+# Core services
+svc bind argus-bind-sys
+svc master argus-master-sys
+svc es argus-es-sys
+svc kibana argus-kibana-sys
+svc ftp argus-ftp
+svc prometheus argus-prometheus
+svc grafana argus-grafana
+svc alertmanager argus-alertmanager
+svc web-frontend argus-web-frontend
+svc web-proxy argus-web-proxy
+
+# HTTP checks (host side)
+section HTTP
+logd "ES: $(http_code "http://localhost:${ES_HTTP_PORT:-9200}/_cluster/health")"
+http_body_head "http://localhost:${ES_HTTP_PORT:-9200}/_cluster/health" >> "$DETAILS" 2>&1 || true
+
+logd "Kibana: $(http_code "http://localhost:${KIBANA_PORT:-5601}/api/status")"
+http_body_head "http://localhost:${KIBANA_PORT:-5601}/api/status" >> "$DETAILS" 2>&1 || true
+
+logd "Master readyz: $(http_code "http://localhost:${MASTER_PORT:-32300}/readyz")"
+
+logd "Prometheus: $(http_code "http://localhost:${PROMETHEUS_PORT:-9090}/-/ready")"
+logd "Grafana: $(http_code "http://localhost:${GRAFANA_PORT:-3000}/api/health")"
+http_body_head "http://localhost:${GRAFANA_PORT:-3000}/api/health" >> "$DETAILS" 2>&1 || true
+logd "Alertmanager: $(http_code "http://localhost:${ALERTMANAGER_PORT:-9093}/api/v2/status")"
+
+cors8084=$(header_val -H "Origin: http://localhost:${WEB_PROXY_PORT_8080:-8080}" "http://localhost:${WEB_PROXY_PORT_8084:-8084}/api/v2/status" || true)
+cors8085=$(header_val -H "Origin: http://localhost:${WEB_PROXY_PORT_8080:-8080}" "http://localhost:${WEB_PROXY_PORT_8085:-8085}/api/v1/master/nodes" || true)
+logd "Web-Proxy 8080: $(http_code "http://localhost:${WEB_PROXY_PORT_8080:-8080}/")"
+logd "Web-Proxy 8083: $(http_code "http://localhost:${WEB_PROXY_PORT_8083:-8083}/")"
+logd "Web-Proxy 8084 CORS: ${cors8084}"
+logd "Web-Proxy 8085 CORS: ${cors8085}"
+
+# Overlay network diagnostics
+section OVERLAY-NET
+if docker network inspect "${OVERLAY_NET_NAME:-argus-sys-net}" >/dev/null 2>&1; then
+  logd "overlay present: ${OVERLAY_NET_NAME:-argus-sys-net}"
+  docker network inspect "${OVERLAY_NET_NAME:-argus-sys-net}" | sed -n '1,60p' >> "$DETAILS" 2>/dev/null || true
+else
+  append_err "[overlay][network] missing ${OVERLAY_NET_NAME:-argus-sys-net}"
+fi
+
+# Domain resolution & reachability from inside web-proxy (bind-backed)
+section DOMAIN
+for d in master.argus.com ftp.argus.com kibana.argus.com es.log.argus.com prom.argus.com grafana.argus.com; do
+  logd "getent $d (web-proxy):"
+  docker exec argus-web-proxy sh -lc "getent hosts $d || true" >> "$DETAILS" 2>&1 || true
+done
+logd "HTTP (web-proxy): master.readyz=$(docker exec argus-web-proxy sh -lc \"curl -s -o /dev/null -w '%{http_code}' http://master.argus.com:3000/readyz\" 2>/dev/null || echo 000)"
+logd "HTTP (web-proxy): es.health=$(docker exec argus-web-proxy sh -lc \"curl -s -o /dev/null -w '%{http_code}' http://es.log.argus.com:9200/_cluster/health\" 2>/dev/null || echo 000)"
+logd "HTTP (web-proxy): kibana.status=$(docker exec argus-web-proxy sh -lc \"curl -s -o /dev/null -w '%{http_code}' http://kibana.argus.com:5601/api/status\" 2>/dev/null || echo 000)"
+
+# FTP share writability (container perspective)
+section FTP-SHARE
+docker exec argus-ftp sh -lc 'ls -ld /private/argus/ftp /private/argus/ftp/share; test -w /private/argus/ftp/share && echo "write:OK" || echo "write:FAIL"' >> "$DETAILS" 2>&1 || true
+
+# Collect system info for context
+section SYSTEM
+logd "uname -a:"; uname -a >> "$DETAILS"
+logd "docker version:"; docker version --format '{{.Server.Version}}' >> "$DETAILS" 2>&1 || true
+logd "compose ps:"; (cd "$ROOT/compose" && if docker compose version >/dev/null 2>&1; then docker compose -p argus-sys ps; else docker-compose -p argus-sys ps; fi) >> "$DETAILS" 2>&1 || true
+
+section SUMMARY
+# Add HTTP failures and CORS problems to error log with tags
+[[ $(http_code "http://localhost:${ES_HTTP_PORT:-9200}/_cluster/health") != 200 ]] && echo "[es][http] health not 200" >> "$ERRORS"
+kbcode=$(http_code "http://localhost:${KIBANA_PORT:-5601}/api/status"); [[ "$kbcode" != 200 ]] && echo "[kibana][http] /api/status=$kbcode" >> "$ERRORS"
+[[ $(http_code "http://localhost:${MASTER_PORT:-32300}/readyz") != 200 ]] && echo "[master][http] /readyz not 200" >> "$ERRORS"
+[[ $(http_code "http://localhost:${PROMETHEUS_PORT:-9090}/-/ready") != 200 ]] && echo "[prometheus][http] /-/ready not 200" >> "$ERRORS"
+gfcode=$(http_code "http://localhost:${GRAFANA_PORT:-3000}/api/health"); [[ "$gfcode" != 200 ]] && echo "[grafana][http] /api/health=$gfcode" >> "$ERRORS"
+[[ $(http_code "http://localhost:${ALERTMANAGER_PORT:-9093}/api/v2/status") != 200 ]] && echo "[alertmanager][http] /api/v2/status not 200" >> "$ERRORS"
+[[ -z "$cors8084" ]] && echo "[web-proxy][cors] 8084 missing Access-Control-Allow-Origin" >> "$ERRORS"
+[[ -z "$cors8085" ]] && echo "[web-proxy][cors] 8085 missing Access-Control-Allow-Origin" >> "$ERRORS"
+
+# Deduplicate errors
+sort -u -o "$ERRORS" "$ERRORS"
+
+echo "Diagnostic details -> $DETAILS"
+echo "Detected errors   -> $ERRORS"
+
+if [[ "$LOG_DIR" == "$ROOT/logs" ]]; then
+  # maintain latest symlinks when writing under package logs
+  ln -sfn "$(basename "$DETAILS")" "$ROOT/logs/diagnose_details.log" 2>/dev/null || cp "$DETAILS" "$ROOT/logs/diagnose_details.log" 2>/dev/null || true
+  ln -sfn "$(basename "$ERRORS")"  "$ROOT/logs/diagnose_error.log"   2>/dev/null || cp "$ERRORS"  "$ROOT/logs/diagnose_error.log"   2>/dev/null || true
+else
+  echo "Diagnostic details -> $DETAILS"
+  echo "Detected errors   -> $ERRORS"
+fi
+
+exit 0
--- a/deployment/build/templates/scripts/server-install.sh
+++ b/deployment/build/templates/scripts/server-install.sh
@@ -0,0 +1,281 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PKG_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"   # version root
+
+PROJECT_NAME="argus-sys"
+
+log() { echo -e "\033[0;34m[INSTALL]\033[0m $*"; }
+err() { echo -e "\033[0;31m[ERROR ]\033[0m $*" >&2; }
+
+require() { command -v "$1" >/dev/null 2>&1 || { err "missing command: $1"; exit 1; }; }
+
+require docker
+if docker compose version >/dev/null 2>&1; then COMPOSE=(docker compose); else require docker-compose; COMPOSE=(docker-compose); fi
+
+ENV_FILE="$PKG_ROOT/compose/.env"
+ENV_TEMPLATE="$PKG_ROOT/compose/.env.example"
+
+find_free_port() {
+  local prefer="$1"; local start=${2:-20000}; local max=${3:-65000};
+  if ! ss -ltnH 2>/dev/null | awk -v pat=":"$prefer"$" '$4 ~ pat{f=1} END{exit f?0:1}'; then echo "$prefer"; return; fi
+  for ((p=start; p<=max; p++)); do
+    if ! ss -ltnH 2>/dev/null | awk -v pat=":"$p"$" '$4 ~ pat{f=1} END{exit f?0:1}'; then echo "$p"; return; fi
+  done
+  return 1
+}
+
+prepare_env() {
+  if [[ -f "$ENV_FILE" ]]; then log ".env exists, keep as-is"; return; fi
+  [[ -f "$ENV_TEMPLATE" ]] || { err "missing $ENV_TEMPLATE"; exit 1; }
+  cp "$ENV_TEMPLATE" "$ENV_FILE"
+  # overlay 模式下，避免为不同服务分配到同一新端口；保持模板端口不做自动改写
+}
+
+prepare_data_dirs() {
+  if [[ $EUID -ne 0 ]]; then
+    echo -e "\033[1;33m[WARN]\033[0m running as non-root: will not chown data dirs."
+    echo -e "\033[1;33m[WARN]\033[0m If you hit Permission denied, run: sudo $SCRIPT_DIR/server-prepare-dirs.sh"
+    # still ensure basic directories exist (no chown)
+    mkdir -p \
+      "$PKG_ROOT/private/argus/etc" \
+      "$PKG_ROOT/private/argus/log/elasticsearch" \
+      "$PKG_ROOT/private/argus/log/kibana" \
+      "$PKG_ROOT/private/argus/metric/prometheus" \
+      "$PKG_ROOT/private/argus/metric/prometheus/data" \
+      "$PKG_ROOT/private/argus/metric/prometheus/rules" \
+      "$PKG_ROOT/private/argus/metric/grafana" \
+      "$PKG_ROOT/private/argus/metric/grafana/data" \
+      "$PKG_ROOT/private/argus/metric/grafana/logs" \
+      "$PKG_ROOT/private/argus/metric/grafana/plugins" \
+      "$PKG_ROOT/private/argus/metric/grafana/provisioning/datasources" \
+      "$PKG_ROOT/private/argus/metric/grafana/provisioning/dashboards" \
+      "$PKG_ROOT/private/argus/metric/grafana/data/sessions" \
+      "$PKG_ROOT/private/argus/metric/grafana/data/dashboards" \
+      "$PKG_ROOT/private/argus/alert/alertmanager" \
+      "$PKG_ROOT/private/argus/metric/ftp/share"
+    # non-root: relax permissions to avoid container UID mismatch blocking writes
+    chmod -R a+rwx "$PKG_ROOT/private/argus" 2>/dev/null || true
+  fi
+}
+
+ensure_swarm_and_overlay() {
+  local net_name="${OVERLAY_NET_NAME:-argus-sys-net}"
+  # Require swarm active
+  local state
+  state=$(docker info --format '{{.Swarm.LocalNodeState}}' 2>/dev/null || echo "")
+  if [[ "$state" != "active" ]]; then
+    err "Docker Swarm is not active. On this host run:"
+    err "  docker swarm init --advertise-addr <this_host_ip>"
+    exit 1
+  fi
+  # Create attachable overlay if missing
+  if ! docker network inspect "$net_name" >/dev/null 2>&1; then
+    log "creating attachable overlay network: $net_name"
+    docker network create --driver overlay --attachable "$net_name" >/dev/null
+  fi
+}
+
+bootstrap_dns_conf() {
+  local etc_dir="$PKG_ROOT/private/argus/etc"
+  mkdir -p "$etc_dir"
+  local dns_file="$etc_dir/dns.conf"
+  if [[ ! -s "$dns_file" ]]; then
+    # detect host primary IP
+    local host_ip
+    host_ip=$(ip route get 1.1.1.1 2>/dev/null | awk '/src/ {print $7; exit}')
+    [[ -z "$host_ip" ]] && host_ip=$(hostname -I 2>/dev/null | awk '{print $1}')
+    if [[ -n "$host_ip" ]]; then
+      echo "$host_ip" > "$dns_file"
+      log "wrote initial dns.conf with host IP: $host_ip"
+    else
+      err "failed to determine host IP for dns.conf; please edit $dns_file manually"
+    fi
+  fi
+}
+
+load_images() {
+  local tar="$PKG_ROOT/images/all-images.tar.gz"
+  [[ -f "$tar" ]] || { err "missing images tar: $tar"; exit 1; }
+  log "loading images from $(basename "$tar") (may take minutes)"
+  gunzip -c "$tar" | docker load >/dev/null
+}
+
+bring_up() {
+  log "starting services via compose"
+  ensure_swarm_and_overlay
+  bootstrap_dns_conf
+  local ov="$PKG_ROOT/compose/docker-compose.os-compat.override.yml"
+  if [[ ! -f "$ov" ]]; then
+    cat > "$ov" <<'YAML'
+services:
+  bind:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+    tmpfs:
+      - /run/named
+  master:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  es:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  kibana:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  ftp:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  prometheus:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  grafana:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  alertmanager:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+    # ensure runtime path matches container expectation
+    volumes:
+      - ../private/argus/etc:/private/argus/etc
+      - ../private/argus/alert/alertmanager:/alertmanager
+  web-frontend:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+  web-proxy:
+    security_opt: ["label=disable"]
+    userns_mode: "host"
+YAML
+    log "generated OS-compat override: $(basename "$ov")"
+  fi
+  # 仅启动服务端组件，避免误起测试节点（node-a/node-b/test-node/test-gpu-node）
+  local services=(bind master es kibana ftp prometheus grafana alertmanager web-frontend web-proxy)
+  log "services: ${services[*]}"
+  (cd "$PKG_ROOT/compose" && "${COMPOSE[@]}" -p "$PROJECT_NAME" -f docker-compose.yml -f $(basename "$ov") up -d "${services[@]}")
+}
+
+# Post bootstrap container-side fixes that do not require sudo on host.
+post_bootstrap_fixes() {
+  # Kibana: ensure /usr/share/kibana/data is a symlink into mounted path to avoid EACCES
+  if docker ps --format '{{.Names}}' | grep -q '^argus-kibana-sys$'; then
+    docker exec argus-kibana-sys bash -lc '
+      set -e
+      mkdir -p /private/argus/log/kibana && chmod 777 /private/argus/log/kibana || true
+      if [ -d /usr/share/kibana/data ] && [ ! -L /usr/share/kibana/data ]; then rm -rf /usr/share/kibana/data; fi
+      if [ ! -e /usr/share/kibana/data ]; then ln -s /private/argus/log/kibana /usr/share/kibana/data; fi
+    ' >/dev/null 2>&1 || true
+  fi
+  # Elasticsearch: ensure data path points to mounted path and is writable
+  if docker ps --format '{{.Names}}' | grep -q '^argus-es-sys$'; then
+    docker exec argus-es-sys bash -lc '
+      set -e
+      mkdir -p /private/argus/log/elasticsearch && chmod 777 /private/argus/log/elasticsearch || true
+      if [ -d /usr/share/elasticsearch/data ] && [ ! -L /usr/share/elasticsearch/data ]; then rm -rf /usr/share/elasticsearch/data; fi
+      if [ ! -e /usr/share/elasticsearch/data ]; then ln -s /private/argus/log/elasticsearch /usr/share/elasticsearch/data; fi
+    ' >/dev/null 2>&1 || true
+  fi
+  # Bind9: ensure rndc.key exists
+  if docker ps --format '{{.Names}}' | grep -q '^argus-bind-sys$'; then
+    docker exec argus-bind-sys bash -lc '
+      set -e
+      mkdir -p /etc/bind
+      if [ ! -f /etc/bind/rndc.key ]; then rndc-confgen -a -c /etc/bind/rndc.key; fi
+      chmod 644 /etc/bind/rndc.key || true
+    ' >/dev/null 2>&1 || true
+  fi
+}
+
+dns_bootstrap() {
+  log "DNS bootstrap: initializing shared dns.conf and container resolv.conf"
+  local etc_dir="$PKG_ROOT/private/argus/etc"
+  mkdir -p "$etc_dir"
+  # 1) ensure dns.conf exists (fallback to bind IP 172.31.0.2)
+  if [[ ! -s "$etc_dir/dns.conf" ]]; then
+    if echo "172.31.0.2" > "$etc_dir/dns.conf" 2>/dev/null; then
+      log "wrote fallback dns.conf with 172.31.0.2"
+    else
+      # host-side write denied (ownership 1000:1000); write via bind container instead
+      if docker ps --format '{{.Names}}' | grep -q '^argus-bind-sys$'; then
+        docker exec argus-bind-sys sh -lc 'echo 172.31.0.2 > /private/argus/etc/dns.conf && chmod 644 /private/argus/etc/dns.conf' || true
+        log "fallback dns.conf written via bind container"
+      else
+        log "bind not ready; skip writing fallback dns.conf"
+      fi
+    fi
+  fi
+  # 2) wait briefly for bind to copy update-dns.sh into shared etc (bind startup.sh does this)
+  local i=0
+  while [[ ! -x "$etc_dir/update-dns.sh" && $i -lt 20 ]]; do
+    sleep 0.5; ((i++));
+  done
+  if [[ ! -x "$etc_dir/update-dns.sh" ]]; then
+    log "update-dns.sh not present yet; continuing with existing resolv.conf"
+  fi
+  # 3) run update-dns.sh inside key containers so /etc/resolv.conf points to bind
+  local c
+  for c in argus-master-sys argus-es-sys argus-kibana-sys argus-grafana argus-prometheus argus-ftp argus-web-frontend argus-web-proxy argus-alertmanager; do
+    if docker ps --format '{{.Names}}' | grep -q "^${c}$"; then
+      docker exec "$c" sh -lc 'test -x /private/argus/etc/update-dns.sh && /private/argus/etc/update-dns.sh || true' >/dev/null 2>&1 || true
+    fi
+  done
+  # 4) wait for service A-record hint files generated by services (best-effort)
+  local need=( es.log.argus.com kibana.log.argus.com master.argus.com grafana.metric.argus.com prom.metric.argus.com alertmanager.alert.argus.com )
+  local waited=0; local missing=1
+  while (( waited < 15 )); do
+    missing=0
+    for f in "${need[@]}"; do [[ -s "$etc_dir/$f" ]] || { missing=1; break; }; done
+    [[ $missing -eq 0 ]] && break
+    sleep 1; ((waited++))
+  done
+  # 5) reload bind zone (script uses supervisor to restart bind9)
+  if docker ps --format '{{.Names}}' | grep -q '^argus-bind-sys$'; then
+    docker exec argus-bind-sys sh -lc '/usr/local/bin/reload-bind9.sh' >/dev/null 2>&1 || true
+  fi
+  # 6) restart web-proxy once to re-render nginx resolver with latest dns.conf
+  if docker ps --format '{{.Names}}' | grep -q '^argus-web-proxy$'; then
+    docker restart argus-web-proxy >/dev/null 2>&1 || true
+  fi
+}
+
+selfcheck() {
+  # Initial selfcheck with retries to absorb cold starts
+  local max_retries="${SELF_CHECK_RETRIES:-5}"   # 重试次数（不含首次），默认 5
+  local wait_seconds="${SELF_CHECK_WAIT_SECONDS:-30}"  # 每次重试前等待秒数，默认 30s
+
+  local attempt=0
+  while :; do
+    attempt=$((attempt+1))
+    if (( attempt == 1 )); then
+      log "running selfcheck (attempt ${attempt})"
+    else
+      log "running selfcheck (attempt ${attempt}/${max_retries}+1)"
+    fi
+
+    if bash "$PKG_ROOT/scripts/server-selfcheck.sh"; then
+      return 0
+    fi
+
+    # failed
+    if (( attempt > max_retries )); then
+      err "selfcheck failed after ${attempt} attempt(s)"
+      exit 1
+    fi
+    log "selfcheck not ready yet; retrying in ${wait_seconds}s..."
+    sleep "$wait_seconds"
+  done
+}
+
+main() {
+  mkdir -p "$PKG_ROOT/logs"
+  prepare_env
+  prepare_data_dirs
+  load_images
+  bring_up
+  post_bootstrap_fixes
+  dns_bootstrap
+  selfcheck
+  log "install completed. See logs in $PKG_ROOT/logs/"
+}
+
+main "$@"
--- a/deployment/build/templates/scripts/server-prepare-dirs.sh
+++ b/deployment/build/templates/scripts/server-prepare-dirs.sh
@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PKG_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+
+if [[ $EUID -ne 0 ]]; then
+  echo "[PREPARE] This script requires root (sudo)." >&2
+  echo "          Try: sudo $0" >&2
+  exit 1
+fi
+
+ENV_FILE="$PKG_ROOT/compose/.env"
+[[ -f "$ENV_FILE" ]] && set -a && source "$ENV_FILE" && set +a
+UIDV="${ARGUS_BUILD_UID:-1000}"; GIDV="${ARGUS_BUILD_GID:-1000}"
+
+echo "[PREPARE] Using owner ${UIDV}:${GIDV}"
+
+# Core etc and service data dirs (aligned with src/sys/tests/scripts/01_bootstrap.sh)
+mkdir -p \
+  "$PKG_ROOT/private/argus/etc" \
+  "$PKG_ROOT/private/argus/bind" \
+  "$PKG_ROOT/private/argus/master" \
+  "$PKG_ROOT/private/argus/agent" \
+  "$PKG_ROOT/private/argus/log/elasticsearch" \
+  "$PKG_ROOT/private/argus/log/kibana"
+
+# Prometheus
+mkdir -p \
+  "$PKG_ROOT/private/argus/metric/prometheus" \
+  "$PKG_ROOT/private/argus/metric/prometheus/data" \
+  "$PKG_ROOT/private/argus/metric/prometheus/rules" \
+  "$PKG_ROOT/private/argus/metric/prometheus/targets"
+
+# Grafana
+mkdir -p \
+  "$PKG_ROOT/private/argus/metric/grafana" \
+  "$PKG_ROOT/private/argus/metric/grafana/data" \
+  "$PKG_ROOT/private/argus/metric/grafana/logs" \
+  "$PKG_ROOT/private/argus/metric/grafana/plugins" \
+  "$PKG_ROOT/private/argus/metric/grafana/provisioning/datasources" \
+  "$PKG_ROOT/private/argus/metric/grafana/provisioning/dashboards" \
+  "$PKG_ROOT/private/argus/metric/grafana/data/sessions" \
+  "$PKG_ROOT/private/argus/metric/grafana/data/dashboards" \
+  "$PKG_ROOT/private/argus/metric/grafana/config"
+
+# FTP
+mkdir -p "$PKG_ROOT/private/argus/metric/ftp/share"
+
+# Alertmanager
+mkdir -p "$PKG_ROOT/private/argus/alert/alertmanager"
+
+chown -R "$UIDV":"$GIDV" \
+  "$PKG_ROOT/private/argus/etc" \
+  "$PKG_ROOT/private/argus/bind" \
+  "$PKG_ROOT/private/argus/master" \
+  "$PKG_ROOT/private/argus/agent" \
+  "$PKG_ROOT/private/argus/log/elasticsearch" \
+  "$PKG_ROOT/private/argus/log/kibana" \
+  "$PKG_ROOT/private/argus/metric/prometheus" \
+  "$PKG_ROOT/private/argus/metric/grafana" \
+  "$PKG_ROOT/private/argus/metric/ftp" \
+  "$PKG_ROOT/private/argus/alert"
+
+chmod -R g+w "$PKG_ROOT/private/argus/alert" "$PKG_ROOT/private/argus/etc" || true
+
+# Ensure parent directories also owned by runtime user for consistency
+chown "$UIDV":"$GIDV" \
+  "$PKG_ROOT/private/argus" \
+  "$PKG_ROOT/private/argus/log" \
+  "$PKG_ROOT/private/argus/metric" || true
+
+echo "[PREPARE] Done. You can now run server-install.sh"
--- a/deployment/build/templates/scripts/server-selfcheck.sh
+++ b/deployment/build/templates/scripts/server-selfcheck.sh
@@ -0,0 +1,104 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+
+log() { echo -e "\033[0;34m[CHECK]\033[0m $*"; }
+err() { echo -e "\033[0;31m[ERROR]\033[0m $*" >&2; }
+
+ENV_FILE="$ROOT/compose/.env"; [[ -f "$ENV_FILE" ]] && set -a && source "$ENV_FILE" && set +a
+
+wait_http() { local url="$1"; local attempts=${2:-120}; local i=1; while ((i<=attempts)); do curl -fsS "$url" >/dev/null 2>&1 && return 0; echo "[..] waiting $url ($i/$attempts)"; sleep 5; ((i++)); done; return 1; }
+code_for() { curl -s -o /dev/null -w "%{http_code}" "$1" || echo 000; }
+header_val() { curl -s -D - -o /dev/null "$@" | awk -F': ' 'BEGIN{IGNORECASE=1}$1=="Access-Control-Allow-Origin"{gsub("\r","",$2);print $2}'; }
+
+LOG_DIR="$ROOT/logs"
+mkdir -p "$LOG_DIR" || true
+OUT_JSON="$LOG_DIR/selfcheck.json"
+tmp=$(mktemp)
+
+ok=1
+
+log "checking overlay network"
+net_ok=false
+if docker network inspect "${OVERLAY_NET_NAME:-argus-sys-net}" >/dev/null 2>&1; then
+  if docker network inspect "${OVERLAY_NET_NAME:-argus-sys-net}" | grep -q '"Driver": "overlay"'; then net_ok=true; fi
+fi
+[[ "$net_ok" == true ]] || ok=0
+
+log "checking Elasticsearch (via domain inside web-proxy)"
+if docker exec argus-web-proxy sh -lc "curl -fsS http://es.log.argus.com:9200/_cluster/health?wait_for_status=yellow&timeout=1s" >/dev/null 2>&1; then es_ok=true; else es_ok=false; ok=0; fi
+
+log "checking Kibana (via domain inside web-proxy)"
+kb_code=$(docker exec argus-web-proxy sh -lc "curl -s -o /dev/null -w '%{http_code}' http://kibana.log.argus.com:5601/api/status" || echo 000)
+kb_ok=false
+if [[ "$kb_code" == "200" ]]; then body=$(curl -sS "http://localhost:${KIBANA_PORT:-5601}/api/status"); echo "$body" | grep -q '"level":"available"' && kb_ok=true; fi
+[[ "$kb_ok" == true ]] || ok=0
+
+log "checking Master (via domain inside web-proxy)"
+if docker exec argus-web-proxy sh -lc "curl -fsS http://master.argus.com:3000/readyz" >/dev/null 2>&1; then true; else ok=0; fi
+
+log "checking FTP"
+if docker ps --format '{{.Names}}' | grep -q '^argus-ftp$'; then
+  if docker exec argus-ftp sh -lc 'test -w /private/argus/ftp/share'; then ftp_ok=true; else ftp_ok=false; ok=0; fi
+else
+  ftp_ok=false; ok=0;
+fi
+
+log "checking Prometheus"
+wait_http "http://localhost:${PROMETHEUS_PORT:-9090}/-/ready" 60 || ok=0
+
+log "checking Grafana"
+gf_code=$(curl -s -o /dev/null -w "%{http_code}" "http://localhost:${GRAFANA_PORT:-3000}/api/health" || echo 000)
+gf_ok=false; if [[ "$gf_code" == "200" ]]; then body=$(curl -sS "http://localhost:${GRAFANA_PORT:-3000}/api/health"); echo "$body" | grep -q '"database"\s*:\s*"ok"' && gf_ok=true; fi
+[[ "$gf_ok" == true ]] || ok=0
+
+log "checking Alertmanager"
+wait_http "http://localhost:${ALERTMANAGER_PORT:-9093}/api/v2/status" 60 || ok=0
+
+log "checking Web-Proxy"
+p8080=$(code_for "http://localhost:${WEB_PROXY_PORT_8080:-8080}/")
+p8083=$(code_for "http://localhost:${WEB_PROXY_PORT_8083:-8083}/")
+cors8084=$(header_val -H "Origin: http://localhost:${WEB_PROXY_PORT_8080:-8080}" "http://localhost:${WEB_PROXY_PORT_8084:-8084}/api/v2/status" || true)
+cors8085=$(header_val -H "Origin: http://localhost:${WEB_PROXY_PORT_8080:-8080}" "http://localhost:${WEB_PROXY_PORT_8085:-8085}/api/v1/master/nodes" || true)
+wp_ok=true
+# 有些环境首页可能 403，此处接受 200/403
+([[ "$p8080" == 200 || "$p8080" == 403 ]]) || wp_ok=false
+([[ "$p8083" == 200 || "$p8083" == 302 || "$p8083" == 403 ]]) || wp_ok=false
+[[ -n "$cors8084" && -n "$cors8085" ]] || wp_ok=false
+[[ "$wp_ok" == true ]] || ok=0
+
+cat > "$tmp" <<JSON
+{
+  "es": $es_ok,
+  "kibana": $kb_ok,
+  "master_readyz": true,
+  "ftp_share_writable": $ftp_ok,
+  "prometheus": true,
+  "grafana": $gf_ok,
+  "alertmanager": true,
+  "web_proxy": $wp_ok,
+  "overlay_net": $net_ok,
+  "timestamp": "$(date -u +%Y-%m-%dT%H:%M:%SZ)"
+}
+JSON
+
+if ! mv "$tmp" "$OUT_JSON" 2>/dev/null; then
+  # fallback when logs dir not writable (no sudo allowed)
+  OUT_JSON="/tmp/selfcheck_$(date -u +%Y%m%d-%H%M%SZ).json"
+  cp "$tmp" "$OUT_JSON"
+  log "selfcheck.json written to $OUT_JSON (logs dir not writable)"
+fi
+if [[ "$ok" == 1 ]]; then
+  log "selfcheck OK"
+  exit 0
+else
+  err "selfcheck FAILED (see $OUT_JSON)"
+  # If diagnose script exists, run it to collect more details
+  if [[ -x "$SCRIPT_DIR/server-diagnose.sh" ]]; then
+    # run diagnose; it will print the actual timestamped file paths and update 'latest' symlinks
+    "$SCRIPT_DIR/server-diagnose.sh" || true
+  fi
+  exit 1
+fi
--- a/deployment/build/templates/scripts/server-status.sh
+++ b/deployment/build/templates/scripts/server-status.sh
@@ -0,0 +1,28 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+
+PROJECT_NAME="argus-sys"
+
+if docker compose version >/dev/null 2>&1; then COMPOSE=(docker compose); else COMPOSE=(docker-compose); fi
+
+echo "== Containers =="
+(cd "$ROOT/compose" && "${COMPOSE[@]}" -p "$PROJECT_NAME" ps)
+
+echo
+echo "== Key Endpoints =="
+ENV_FILE="$ROOT/compose/.env"; [[ -f "$ENV_FILE" ]] && set -a && source "$ENV_FILE" && set +a
+printf "master    http://localhost:%s/readyz\n" "${MASTER_PORT:-32300}"
+printf "es        http://localhost:%s/_cluster/health\n" "${ES_HTTP_PORT:-9200}"
+printf "kibana    http://localhost:%s/api/status\n" "${KIBANA_PORT:-5601}"
+printf "prom      http://localhost:%s/-/ready\n" "${PROMETHEUS_PORT:-9090}"
+printf "grafana   http://localhost:%s/api/health\n" "${GRAFANA_PORT:-3000}"
+printf "alert     http://localhost:%s/api/v2/status\n" "${ALERTMANAGER_PORT:-9093}"
+printf "web       http://localhost:%s/ (8080)\n" "${WEB_PROXY_PORT_8080:-8080}"
+
+echo
+echo "== Selfcheck result =="
+cat "$ROOT/logs/selfcheck.json" 2>/dev/null || echo "(no selfcheck yet)"
+
--- a/deployment/build/templates/scripts/server-uninstall.sh
+++ b/deployment/build/templates/scripts/server-uninstall.sh
@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PKG_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+
+PROJECT_NAME="argus-sys"
+
+log() { echo -e "\033[0;34m[UNINSTALL]\033[0m $*"; }
+
+if docker compose version >/dev/null 2>&1; then COMPOSE=(docker compose); else COMPOSE=(docker-compose); fi
+
+(cd "$PKG_ROOT/compose" && "${COMPOSE[@]}" -p "$PROJECT_NAME" down -v || true)
+log "compose stack removed"
+log "you may remove data under $PKG_ROOT/private if you want a clean slate"
+
--- a/doc/metric_lists.xlsx
+++ b/doc/metric_lists.xlsx
--- a/src/agent/scripts/build_binary.sh
+++ b/src/agent/scripts/build_binary.sh
@@ -12,6 +12,8 @@ VENV_DIR="$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
--- a/src/alert/alertmanager/build/Dockerfile
+++ b/src/alert/alertmanager/build/Dockerfile
@@ -9,21 +9,21 @@ 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
+# 设置 Alertmanager 版本（与本地离线包保持一致）
 ARG ALERTMANAGER_VERSION=0.28.1

-# 下载并解压 Alertmanager 二进制
-RUN wget https://github.com/prometheus/alertmanager/releases/download/v${ALERTMANAGER_VERSION}/alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz && \
-    tar xvf alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz && \
-    mv alertmanager-${ALERTMANAGER_VERSION}.linux-amd64 /usr/local/alertmanager && \
-    rm alertmanager-${ALERTMANAGER_VERSION}.linux-amd64.tar.gz
+# 使用仓库内预置的离线包构建（无需联网）
+COPY src/alert/alertmanager/build/alertmanager-${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_UID=2133
-ARG ARGUS_GID=2015
-ENV ARGUS_UID=${ARGUS_UID}
-ENV ARGUS_GID=${ARGUS_GID}
+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} && \
@@ -33,16 +33,24 @@ RUN mkdir -p /usr/share/alertmanager && \

 # 创建 alertmanager 用户（可自定义 UID/GID）
 # 创建 alertmanager 用户组
-RUN groupadd -g ${ARGUS_GID} alertmanager
+RUN set -eux; \
+    # 确保目标 GID 存在；若已被占用，直接使用该 GID（组名不限）\
+    if ! getent group "${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

-# 创建 alertmanager 用户并指定组
-RUN useradd -M -s /usr/sbin/nologin -u ${ARGUS_UID} -g ${ARGUS_GID} alertmanager
-
-RUN chown -R alertmanager:alertmanager /usr/share/alertmanager && \
-    chown -R alertmanager:alertmanager /alertmanager && \
-    chown -R alertmanager:alertmanager ${ALERTMANAGER_BASE_PATH} && \
-    chown -R alertmanager:alertmanager /private/argus/etc && \
-    chown -R alertmanager:alertmanager /usr/local/bin
+RUN chown -R "${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 \
@@ -86,4 +94,3 @@ EXPOSE 9093

 # 使用 supervisor 作为入口点
 CMD ["/usr/bin/supervisord", "-c", "/etc/supervisor/conf.d/supervisord.conf"]
-
--- a/src/alert/alertmanager/build/alertmanager-0.28.1.linux-amd64.tar.gz
+++ b/src/alert/alertmanager/build/alertmanager-0.28.1.linux-amd64.tar.gz
--- a/src/alert/alertmanager/build/build.sh
+++ b/src/alert/alertmanager/build/build.sh
@@ -5,8 +5,8 @@ docker pull ubuntu:24.04
 source src/alert/tests/.env

 docker build \
-  --build-arg ARGUS_UID=${ARGUS_UID} \
-  --build-arg ARGUS_GID=${ARGUS_GID} \
+  --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 .

--- a/src/alert/alertmanager/build/fetch-dist.sh
+++ b/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"
+
--- a/src/alert/alertmanager/build/start-am-supervised.sh
+++ b/src/alert/alertmanager/build/start-am-supervised.sh
@@ -7,10 +7,8 @@ ALERTMANAGER_BASE_PATH=${ALERTMANAGER_BASE_PATH:-/private/argus/alert/alertmanag

 echo "[INFO] Alertmanager base path: ${ALERTMANAGER_BASE_PATH}"

-# 生成配置文件
-echo "[INFO] Generating Alertmanager configuration file..."
-sed "s|\${ALERTMANAGER_BASE_PATH}|${ALERTMANAGER_BASE_PATH}|g" \
-    /etc/alertmanager/alertmanager.yml > ${ALERTMANAGER_BASE_PATH}/alertmanager.yml
+# 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
--- a/src/alert/alertmanager/build/supervisord.conf
+++ b/src/alert/alertmanager/build/supervisord.conf
@@ -6,7 +6,7 @@ user=root

 [program:alertmanager]
 command=/usr/local/bin/start-am-supervised.sh
-user=alertmanager
+user=ubuntu
 stdout_logfile=/var/log/supervisor/alertmanager.log
 stderr_logfile=/var/log/supervisor/alertmanager_error.log
 autorestart=true
--- a/src/alert/tests/.env
+++ b/src/alert/tests/.env
@@ -1,5 +1,5 @@
 DATA_ROOT=/home/argus/tmp/private/argus
-ARGUS_UID=1048
-ARGUS_GID=1048
+ARGUS_BUILD_UID=1048
+ARGUS_BUILD_GID=1048

 USE_INTRANET=false
--- a/src/alert/tests/.env.example
+++ b/src/alert/tests/.env.example
@@ -1,5 +1,5 @@
 DATA_ROOT=/home/argus/tmp/private/argus
-ARGUS_UID=1048
-ARGUS_GID=1048
+ARGUS_BUILD_UID=1048
+ARGUS_BUILD_GID=1048

 USE_INTRANET=false
--- a/src/alert/tests/data/alertmanager/alertmanager.yml
+++ b/src/alert/tests/data/alertmanager/alertmanager.yml
@@ -1,19 +0,0 @@
-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']
--- a/src/alert/tests/data/alertmanager/nflog
+++ b/src/alert/tests/data/alertmanager/nflog
--- a/src/alert/tests/data/alertmanager/silences
+++ b/src/alert/tests/data/alertmanager/silences
--- a/src/alert/tests/data/etc/alertmanager.alert.argus.com
+++ b/src/alert/tests/data/etc/alertmanager.alert.argus.com
@@ -1 +0,0 @@
-172.18.0.2
--- a/src/alert/tests/docker-compose.yml
+++ b/src/alert/tests/docker-compose.yml
@@ -4,15 +4,15 @@ services:
      context: ../../../
      dockerfile: src/alert/alertmanager/build/Dockerfile
      args:
-        ARGUS_UID: ${ARGUS_UID:-2133}
-        ARGUS_GID: ${ARGUS_GID:-2015}
+        ARGUS_BUILD_UID: ${ARGUS_BUILD_UID:-2133}
+        ARGUS_BUILD_GID: ${ARGUS_BUILD_GID:-2015}
        USE_INTRANET: ${USE_INTRANET:-false}
    image: argus-alertmanager:latest
    container_name: argus-alertmanager
    environment:
      - ALERTMANAGER_BASE_PATH=/private/argus/alert/alertmanager
-      - ARGUS_UID=${ARGUS_UID:-2133}
-      - ARGUS_GID=${ARGUS_GID:-2015}
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
    ports:
      - "${ARGUS_PORT:-9093}:9093"
    volumes:
--- a/src/alert/tests/scripts/01_bootstrap.sh
+++ b/src/alert/tests/scripts/01_bootstrap.sh
@@ -1,19 +0,0 @@
-#!/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"
--- a/src/alert/tests/scripts/02_up.sh
+++ b/src/alert/tests/scripts/02_up.sh
@@ -1,10 +0,0 @@
-#!/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"
--- a/src/alert/tests/scripts/03_alertmanager_add_alert.sh
+++ b/src/alert/tests/scripts/03_alertmanager_add_alert.sh
@@ -1,106 +0,0 @@
-#!/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
--- a/src/alert/tests/scripts/04_query_alerts.sh
+++ b/src/alert/tests/scripts/04_query_alerts.sh
@@ -1,71 +0,0 @@
-#!/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"
--- a/src/alert/tests/scripts/05_down.sh
+++ b/src/alert/tests/scripts/05_down.sh
@@ -1,21 +0,0 @@
-#!/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 "$0")/.."
-if [ -d "private" ]; then
-    # 删除private目录及其所有内容
-    rm -rf private
-    echo "[OK] 已清理private目录"
-else
-    echo "[INFO] private目录不存在，无需清理"
-fi
--- a/src/alert/tests/scripts/e2e_test.sh
+++ b/src/alert/tests/scripts/e2e_test.sh
@@ -1,105 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-echo "======================================="
-echo "ARGUS Alert System End-to-End Test"
-echo "======================================="
-echo ""
-
-# 记录测试开始时间
-test_start_time=$(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: $am_status"
-echo "   • All services started and stopped successfully"
-echo ""
-echo "✅ The ARGUS Alert system is working correctly!"
-echo ""
--- a/src/alert/tests/scripts/verify_alertmanager.sh
+++ b/src/alert/tests/scripts/verify_alertmanager.sh
@@ -0,0 +1,113 @@
+#!/bin/bash
+# verify_alertmanager.sh
+# 用于部署后验证 Prometheus 与 Alertmanager 通信链路是否正常
+
+set -euo pipefail
+
+#=============================
+# 基础配置
+#=============================
+PROM_URL="${PROM_URL:-http://prom.metric.argus.com:9090}"
+ALERT_URL="${ALERT_URL:-http://alertmanager.alert.argus.com:9093}"
+# TODO: 根据实际部署环境调整规则目录
+DATA_ROOT="${DATA_ROOT:-/private/argus}"
+RULE_DIR = "$DATA_ROOT/metric/prometheus/rules"
+TMP_RULE="/tmp/test_rule.yml"
+
+#=============================
+# 辅助函数
+#=============================
+GREEN="\033[32m"; RED="\033[31m"; YELLOW="\033[33m"; RESET="\033[0m"
+
+log_info()    { echo -e "${YELLOW}[INFO]${RESET} $1"; }
+log_success() { echo -e "${GREEN}[OK]${RESET} $1"; }
+log_error()   { echo -e "${RED}[ERROR]${RESET} $1"; }
+
+fail_exit() { log_error "$1"; exit 1; }
+
+#=============================
+# Step 1: 检查 Alertmanager 是否可访问
+#=============================
+log_info "检查 Alertmanager 状态..."
+if curl -sSf "${ALERT_URL}/api/v2/status" >/dev/null 2>&1; then
+  log_success "Alertmanager 服务正常 (${ALERT_URL})"
+else
+  fail_exit "无法访问 Alertmanager，请检查端口映射与容器状态。"
+fi
+
+#=============================
+# Step 2: 手动发送测试告警
+#=============================
+log_info "发送手动测试告警..."
+curl -s -XPOST "${ALERT_URL}/api/v2/alerts" -H "Content-Type: application/json" -d '[
+  {
+    "labels": {
+      "alertname": "ManualTestAlert",
+      "severity": "info"
+    },
+    "annotations": {
+      "summary": "This is a test alert from deploy verification"
+    },
+    "startsAt": "'$(date -Iseconds)'"
+  }
+]' >/dev/null && log_success "测试告警已成功发送到 Alertmanager"
+
+#=============================
+# Step 3: 检查 Prometheus 配置中是否包含 Alertmanager
+#=============================
+log_info "检查 Prometheus 是否配置了 Alertmanager..."
+if curl -s "${PROM_URL}/api/v1/status/config" | grep -q "alertmanagers"; then
+  log_success "Prometheus 已配置 Alertmanager 目标"
+else
+  fail_exit "Prometheus 未配置 Alertmanager，请检查 prometheus.yml"
+fi
+
+#=============================
+# Step 4: 创建并加载测试告警规则
+#=============================
+log_info "创建临时测试规则 ${TMP_RULE} ..."
+cat <<EOF > "${TMP_RULE}"
+groups:
+- name: deploy-verify-group
+  rules:
+  - alert: DeployVerifyAlert
+    expr: vector(1)
+    labels:
+      severity: warning
+    annotations:
+      summary: "Deployment verification alert"
+EOF
+
+mkdir -p "${RULE_DIR}"
+cp "${TMP_RULE}" "${RULE_DIR}/test_rule.yml"
+
+log_info "重载 Prometheus 以加载新规则..."
+if curl -s -X POST "${PROM_URL}/-/reload" >/dev/null; then
+  log_success "Prometheus 已重载规则"
+else
+  fail_exit "Prometheus reload 失败，请检查 API 可访问性。"
+fi
+
+#=============================
+# Step 5: 等待并验证 Alertmanager 是否收到告警
+#=============================
+log_info "等待告警触发 (约5秒)..."
+sleep 5
+
+if curl -s "${ALERT_URL}/api/v2/alerts" | grep -q "DeployVerifyAlert"; then
+  log_success "Prometheus → Alertmanager 告警链路验证成功"
+else
+  fail_exit "未在 Alertmanager 中检测到 DeployVerifyAlert，请检查网络或配置。"
+fi
+
+#=============================
+# Step 6: 清理测试规则
+#=============================
+log_info "清理临时测试规则..."
+rm -f "${RULE_DIR}/test_rule.yml" "${TMP_RULE}"
+
+curl -s -X POST "${PROM_URL}/-/reload" >/dev/null \
+  && log_success "Prometheus 已清理验证规则" \
+  || log_error "Prometheus reload 清理失败，请手动确认。"
+
+log_success "部署验证全部通过！Prometheus ↔ Alertmanager 通信正常。"
--- a/src/bind/build/Dockerfile
+++ b/src/bind/build/Dockerfile
@@ -26,6 +26,7 @@ RUN apt-get update && \
    apt-get install -y \
    bind9 \
    bind9utils \
+    dnsutils \
    bind9-doc \
    supervisor \
    net-tools \
--- a/src/log/fluent-bit/build/packages/libbrotli1_1.0.9-2build6_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libbrotli1_1.0.9-2build6_amd64.deb
--- a/src/log/fluent-bit/build/packages/libidn2-0_2.3.2-2build1_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libidn2-0_2.3.2-2build1_amd64.deb
--- a/src/log/fluent-bit/build/packages/libldap-2.5-0_2.5.19+dfsg-0ubuntu0.22.04.1_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libldap-2.5-0_2.5.19+dfsg-0ubuntu0.22.04.1_amd64.deb
--- a/src/log/fluent-bit/build/packages/libpq5_14.19-0ubuntu0.22.04.1_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libpq5_14.19-0ubuntu0.22.04.1_amd64.deb
--- a/src/log/fluent-bit/build/packages/libsasl2-2_2.1.27+dfsg2-3ubuntu1.2_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libsasl2-2_2.1.27+dfsg2-3ubuntu1.2_amd64.deb
--- a/src/log/fluent-bit/build/packages/libsasl2-modules-db_2.1.27+dfsg2-3ubuntu1.2_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libsasl2-modules-db_2.1.27+dfsg2-3ubuntu1.2_amd64.deb
--- a/src/log/fluent-bit/build/packages/libssl3_3.0.2-0ubuntu1.20_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libssl3_3.0.2-0ubuntu1.20_amd64.deb
--- a/src/log/fluent-bit/build/packages/libyaml-0-2_0.2.2-1build2_amd64.deb
+++ b/src/log/fluent-bit/build/packages/libyaml-0-2_0.2.2-1build2_amd64.deb
--- a/src/log/fluent-bit/build/start-fluent-bit.sh
+++ b/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
--- a/src/log/tests/scripts/01_bootstrap.sh
+++ b/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
--- a/src/metric/.gitignore
+++ b/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/
--- a/src/metric/client-plugins/all-in-one-demo/scripts/publish_artifact.sh
+++ b/src/metric/client-plugins/all-in-one-demo/scripts/publish_artifact.sh
@@ -104,7 +104,26 @@ 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: $PUBLISH_DIR"
-sudo mkdir -p "$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-$$"
@@ -208,26 +227,31 @@ fi
 TAR_NAME="argus-metric_$(echo $VERSION | tr '.' '_').tar.gz"
 log_info "创建发布包: $TAR_NAME"
 cd "$TEMP_PACKAGE_DIR"
-sudo tar -czf "$PUBLISH_DIR/$TAR_NAME" *
+tar -czf "$PUBLISH_DIR/$TAR_NAME" *
 cd - > /dev/null

-# 设置文件所有者
-log_info "设置文件所有者为: $OWNER"
-sudo chown "$OWNER" "$PUBLISH_DIR/$TAR_NAME"
+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 "$VERSION" | sudo tee "$PUBLISH_DIR/LATEST_VERSION" > /dev/null
-sudo chown "$OWNER" "$PUBLISH_DIR/LATEST_VERSION"
+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..."
-    sudo cp "config/dns.conf" "$PUBLISH_DIR/"
-    sudo chown "$OWNER" "$PUBLISH_DIR/dns.conf"
+    cp "config/dns.conf" "$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: $PUBLISH_DIR/dns.conf"
 else
    log_warning "未找到 config/dns.conf 文件，跳过 DNS 配置文件复制"
@@ -236,8 +260,10 @@ fi
 # 复制 setup.sh 到发布目录
 if [[ -f "scripts/setup.sh" ]]; then
    log_info "复制 setup.sh 到发布目录..."
-    sudo cp "scripts/setup.sh" "$PUBLISH_DIR/"
-    sudo chown "$OWNER" "$PUBLISH_DIR/setup.sh"
+    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
--- a/src/metric/client-plugins/all-in-one-full/README.md
+++ b/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。
--- a/src/metric/client-plugins/all-in-one-full/config/.VERSION.example
+++ b/src/metric/client-plugins/all-in-one-full/config/.VERSION.example
@@ -0,0 +1 @@
+1.29.0
--- a/src/metric/client-plugins/all-in-one-full/config/.checklist.example
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/config/VERSION
+++ b/src/metric/client-plugins/all-in-one-full/config/VERSION
@@ -0,0 +1 @@
+1.40.0
--- a/src/metric/client-plugins/all-in-one-full/config/checklist
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/config/config.env
+++ b/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!
--- a/src/metric/client-plugins/all-in-one-full/config/config.env.example
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/config/dns.conf
+++ b/src/metric/client-plugins/all-in-one-full/config/dns.conf
@@ -0,0 +1 @@
+172.31.0.2
--- a/src/metric/client-plugins/all-in-one-full/deps/cron-offline.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/cron-offline.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/jq-curl.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/jq-curl.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/cron.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/cron.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/curl.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/curl.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/jq.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu20/jq.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/cron.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/cron.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/curl.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/curl.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/jq.tar.gz
+++ b/src/metric/client-plugins/all-in-one-full/deps/ubuntu22/jq.tar.gz
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/README.md
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/bin/argus-agent
+++ b/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/bin/argus-agent
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/check_health.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/install.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/package.sh
+++ b/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"
--- a/src/metric/client-plugins/all-in-one-full/plugins/argus-agent/uninstall.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/bin/datacenter-gpu-manager_3.3.9_amd64.deb
+++ b/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/bin/datacenter-gpu-manager_3.3.9_amd64.deb
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/bin/dcgm-exporter
+++ b/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/bin/dcgm-exporter
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/check_health.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/config/default-counters.csv
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/install.sh
+++ b/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 "$@"
+fi
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/package.sh
+++ b/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/package.sh
@@ -0,0 +1,97 @@
+#!/bin/bash
+
+set -e
+
+# 颜色定义
+GREEN='\033[0;32m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+log_info() {
+    echo -e "${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 "  - $file"
+    done
+    exit 1
+fi
+
+# 防御：阻止将 Git LFS 指针文件打包
+for f in bin/dcgm-exporter bin/datacenter-gpu-manager_3.3.9_amd64.deb; do
+  if head -n1 "$f" 2>/dev/null | grep -q '^version https://git-lfs.github.com/spec/v1$'; then
+      echo "[ERROR] $f 是 Git LFS 指针文件，未还原为真实制品"
+      echo "        请在仓库根目录执行: git lfs fetch --all && git lfs checkout"
+      exit 1
+  fi
+done
+
+log_success "所有必要文件检查完成"
+
+# 创建临时目录
+TEMP_DIR=$(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"
--- a/src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/uninstall.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/README.md
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/fluent-bit_3.1.9_amd64.deb
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/fluent-bit_3.1.9_amd64.deb
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libpq5_14.19-0ubuntu0.22.04.1_amd64.deb
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libpq5_14.19-0ubuntu0.22.04.1_amd64.deb
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libyaml-0-2_0.2.2-1build2_amd64.deb
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/bin/libyaml-0-2_0.2.2-1build2_amd64.deb
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/check_health.sh
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/fluent-bit.conf
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inject_labels.lua
+++ b/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
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inputs.d/10-train.conf
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inputs.d/10-train.conf
@@ -0,0 +1,10 @@
+[INPUT]
+    Name              tail
+    Path              /logs/train/*.log
+    Tag               app.train
+    Path_Key          log_path
+    Refresh_Interval  5
+    DB                /buffers/train.db
+    Skip_Long_Lines   On
+    storage.type      filesystem
+    multiline.parser  python,go,java
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inputs.d/20-infer.conf
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/inputs.d/20-infer.conf
@@ -0,0 +1,10 @@
+[INPUT]
+    Name              tail
+    Path              /logs/infer/*.log
+    Tag               app.infer
+    Path_Key          log_path
+    Refresh_Interval  5
+    DB                /buffers/infer.db
+    Skip_Long_Lines   On
+    storage.type      filesystem
+    multiline.parser  python,go,java
--- a/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/outputs.d/10-es.conf
+++ b/src/metric/client-plugins/all-in-one-full/plugins/fluent-bit/config/outputs.d/10-es.conf
@@ -0,0 +1,24 @@
+# 重要：使用 Logstash_Format + Logstash_Prefix，生成 train-*/infer-* 索引
+[OUTPUT]
+    Name                es
+    Match               app.train
+    Host                ${ES_HOST:-localhost}
+    Port                ${ES_PORT:-9200}
+    Logstash_Format     On
+    Logstash_Prefix     train
+    Replace_Dots        On
+    Generate_ID         On
+    Retry_Limit         False
+    Suppress_Type_Name  On
+
+[OUTPUT]
+    Name                es
+    Match               app.infer
+    Host                ${ES_HOST:-localhost}
+    Port                ${ES_PORT:-9200}
+    Logstash_Format     On
+    Logstash_Prefix     infer
+    Replace_Dots        On
+    Generate_ID         On
+    Retry_Limit         False
+    Suppress_Type_Name  On
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
yuyr	2ff7c55f3b	[#37 ] 测试通过swarm跨机部署节点；更新文档	2025-11-05 09:57:08 +08:00
yuyr	9858f4471e	[#37 ] server install 增加重试自检	2025-11-05 09:57:08 +08:00
yuyr	c8279997a4	[#37 ] swarm 部署优化	2025-11-05 09:57:08 +08:00
yuyr	4ed5c64804	[#37 ] 优化client构建	2025-11-05 09:57:08 +08:00
yuyr	3551360687	[#37 ] 优化client安装包	2025-11-05 09:57:08 +08:00
yuyr	3202e02b42	[#37 ] 测试NixOS部署server通过	2025-11-05 09:57:07 +08:00
yuyr	29eb75a374	[#37 ] 构建安装包	2025-11-05 09:57:07 +08:00
yuyr	ccc141f557	[#30 ] ftp容器增加动态检测并更新dns.conf到share目录	2025-11-05 09:57:07 +08:00
yuyr	ed0d1ca904	[#30 ] wsl部署测试；更新README	2025-11-05 09:57:07 +08:00
xuxt	b6da5bc8b8	dev_1.0.0_xuxt_3 完成web和alert的集成测试 (#38 ) Co-authored-by: xiuting.xu <xiutingxt.xu@gmail.com> Reviewed-on: #38 Reviewed-by: huhy <husteryezi@163.com> Reviewed-by: yuyr <yuyr@zgclab.edu.cn> Reviewed-by: sundapeng <sundp@mail.zgclab.edu.cn>	2025-10-31 14:18:19 +08:00
yuyr	59a38513a4	完成a6000测试系统构建、部署、测试整合 (#35 ) 测试方案： - lm2机器端口映射到本机：18080, 18081, 8082-8085 - 访问URL: http://localhost:18080/dashboard ![image.png](/attachments/30ed6e20-697a-4d3b-a6d3-6acccd2e9922) ![image.png](/attachments/38ef1751-0f3b-49c6-9100-f70d15617acc) ![image.png](/attachments/3be45005-9b9e-4165-8ef6-1d27405800f1) ![image.png](/attachments/eb916192-edc1-4096-8f9f-9769ab6d9039) ![image.png](/attachments/620e6efc-bd02-45ae-bba1-99a95a1b4c02) ![image.png](/attachments/986e77e7-c687-405f-a760-93282249f72f) 端到端测试通过： ![image.png](/attachments/c6e29875-4a16-4718-8b2f-368f64eb545e) Co-authored-by: sundapeng.sdp <sundapeng@hashdata.cn> Reviewed-on: #35 Reviewed-by: xuxt <xuxt@zgclab.edu.cn> Reviewed-by: sundapeng <sundp@mail.zgclab.edu.cn> Reviewed-by: huhy <husteryezi@163.com>	2025-10-29 10:04:27 +08:00
xuxt	d1b89c0cf6	dev_1.0.0_xuxt_2 更新反向代理，打包镜像，以及README文档 (#28 ) Co-authored-by: xiuting.xu <xiutingxt.xu@gmail.com> Reviewed-on: #28 Reviewed-by: yuyr <yuyr@zgclab.edu.cn> Reviewed-by: huhy <husteryezi@163.com> Reviewed-by: sundapeng <sundp@mail.zgclab.edu.cn>	2025-10-20 09:45:32 +08:00
				`@@ -0,0 +1 @@`
				`src/metric/client-plugins/all-in-one-full/plugins//bin/ filter=lfs diff=lfs merge=lfs -text`