fix: support preproduction monitoring on arm64

docs/production-deployment.md

@@ -75,7 +75,7 @@ PROD_ADMIN_USERNAME=prod_admin
 PROD_ADMIN_PASSWORD='change-me'
 ```
 
-系统监控需要在发布前单独安装，配置和规则位于`tools/monitoring/`。在Debian/Ubuntu服务器依次执行`bash tools/monitoring/install-prometheus-monitoring.sh`和`bash tools/monitoring/install-service-exporters.sh`；前者安装Prometheus/Node Exporter，后者安装PostgreSQL/Redis/Nginx Exporter并开启MinIO回环原生指标。脚本必须先备份现有配置并运行`promtool`校验；9090、9100、9187、9121、9113、9464和API/Gateway控制端口必须只监听`127.0.0.1`，不得加入Nginx公网反向代理或安全组放行。安装完成且确认全部target为up后，才可完成发布；详细口径见`docs/prometheus-system-monitoring-design-20260814.md`。
+系统监控需要在发布前单独安装，配置和规则位于`tools/monitoring/`。在 Debian/Ubuntu（apt）或 Alibaba Cloud Linux 4 ARM64 服务器依次执行`bash tools/monitoring/install-prometheus-monitoring.sh`和`bash tools/monitoring/install-service-exporters.sh`；apt 环境使用系统包，Alibaba Cloud Linux 4 ARM64 使用脚本内锁定版本与 SHA-256 的官方 release 二进制。前者安装 Prometheus/Node Exporter，后者安装 PostgreSQL/Redis/Nginx Exporter 并开启 MinIO 回环原生指标。脚本必须先备份已有配置并运行`promtool`校验；9090、9100、9187、9121、9113、9464和API/Gateway控制端口必须只监听`127.0.0.1`，不得加入Nginx公网反向代理或安全组放行。安装完成且确认全部 target 为 up 后，才可完成发布；详细口径见`docs/prometheus-system-monitoring-design-20260814.md`。
 
 安全会话使用 HttpOnly Cookie，正式生产必须先为页面和管理 API 配置 HTTPS，并保持 `SESSION_COOKIE_SECURE=true`；纯 HTTP 的 `IP:12026` 不作为受支持的登录入口，即使切换期仍保留其监听，也只允许用于非登录的兼容检查并应尽快下线。`sms.lisglo.com` 只允许 Cloudflare 回源，`api.lisglo.com` 通过独立 Nginx SNI 虚拟主机只开放客户接口、客户 Swagger 和健康检查；Let’s Encrypt 使用 DNS-01 自动续期，不依赖开放 80 端口。
 

tools/monitoring/README.md

@@ -2,7 +2,7 @@
 
 本目录提供系统监控第一版所需的真实采集配置。运营端不嵌入Prometheus或Grafana页面；NestJS仅从本机Prometheus读取固定指标，再由平台原生UI展示。
 
-在目标Debian/Ubuntu测试服务器上，以root执行：
+在目标 Debian/Ubuntu（apt）或 Alibaba Cloud Linux 4 ARM64 测试服务器上，以 root 执行：
 
 ```bash
 cd /opt/cmpp-platform
@@ -10,7 +10,7 @@ bash tools/monitoring/install-prometheus-monitoring.sh
 bash tools/monitoring/install-service-exporters.sh
 ```
 
-第一个脚本安装Prometheus与Node Exporter，备份已有配置并校验规则；第二个脚本安装PostgreSQL、Redis和Nginx Exporter，开启MinIO回环原生指标。API指标仅监听`127.0.0.1:9464`，Gateway指标复用回环控制端口`8090`。9090、9100、9187、9121、9113和9464均不得对公网开放。两个脚本均会打印恢复资产路径。
+Debian/Ubuntu 使用系统包；Alibaba Cloud Linux 4 ARM64 使用脚本中锁定版本和 SHA-256 的官方 release 二进制。第一个脚本安装 Prometheus 与 Node Exporter，备份已有配置并校验规则；第二个脚本安装 PostgreSQL、Redis 和 Nginx Exporter，开启 MinIO 回环原生指标。API 指标仅监听 `127.0.0.1:9464`，Gateway 指标复用回环控制端口 `8090`。9090、9100、9187、9121、9113 和 9464 均不得对公网开放。两个脚本均会打印恢复资产路径。
 
 安装后将下列配置写入`/etc/cmpp-platform/cmpp-platform.env`，再按正常发布窗口重启API：
 

tools/monitoring/install-prometheus-monitoring.sh

@@ -4,27 +4,113 @@ set -Eeuo pipefail
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 PROMETHEUS_RETENTION_TIME="${PROMETHEUS_RETENTION_TIME:-30d}"
 PROMETHEUS_RETENTION_SIZE="${PROMETHEUS_RETENTION_SIZE:-8GB}"
+PROMETHEUS_VERSION="3.14.0"
+PROMETHEUS_SHA256_ARM64="077f3781ab7245dc04c9a3c9b78ba120fc8e41aa0dc97489b0af67247e50ba83"
+NODE_EXPORTER_VERSION="1.12.1"
+NODE_EXPORTER_SHA256_ARM64="ad35b605f9954b9f1ffddf5ba054bdc5a98d790b9eae5291e1eeb83f1ecbd0e7"
 
 if [[ "$(id -u)" -ne 0 ]]; then
   echo "Run as root." >&2
   exit 1
 fi
-if ! command -v apt-get >/dev/null 2>&1; then
-  echo "This installer currently supports Debian/Ubuntu apt packages only." >&2
-  exit 1
-fi
-
 log() { printf '\n[%s] %s\n' "$(date '+%F %T')" "$*"; }
 
-log "Installing Prometheus and Node Exporter packages"
+install_release_binary() {
-apt-get update
+  local url="$1" expected_sha256="$2" archive_member="$3" destination="$4"
-DEBIAN_FRONTEND=noninteractive apt-get install -y prometheus prometheus-node-exporter curl iproute2
+  local work_dir archive
+  work_dir="$(mktemp -d /tmp/cmpp-monitoring.XXXXXX)"
+  archive="$work_dir/release.tar.gz"
+  curl -fL --retry 5 --retry-delay 2 --connect-timeout 15 -o "$archive" "$url"
+  printf '%s  %s\n' "$expected_sha256" "$archive" | sha256sum -c -
+  tar -xzf "$archive" -C "$work_dir"
+  install -o root -g root -m 0755 "$work_dir/$archive_member" "$destination"
+  rm -rf -- "$work_dir"
+}
+
+install_official_arm64_releases() {
+  [[ "$(uname -m)" == "aarch64" ]] || {
+    echo "Official-binary installation currently supports Linux aarch64 only." >&2
+    exit 1
+  }
+  for command_name in curl sha256sum tar; do
+    command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }
+  done
+
+  log "Installing checksum-pinned official Prometheus ARM64 releases"
+  install_release_binary \
+    "https://github.com/prometheus/prometheus/releases/download/v${PROMETHEUS_VERSION}/prometheus-${PROMETHEUS_VERSION}.linux-arm64.tar.gz" \
+    "$PROMETHEUS_SHA256_ARM64" \
+    "prometheus-${PROMETHEUS_VERSION}.linux-arm64/prometheus" \
+    /usr/local/bin/prometheus
+  install_release_binary \
+    "https://github.com/prometheus/prometheus/releases/download/v${PROMETHEUS_VERSION}/prometheus-${PROMETHEUS_VERSION}.linux-arm64.tar.gz" \
+    "$PROMETHEUS_SHA256_ARM64" \
+    "prometheus-${PROMETHEUS_VERSION}.linux-arm64/promtool" \
+    /usr/local/bin/promtool
+  install_release_binary \
+    "https://github.com/prometheus/node_exporter/releases/download/v${NODE_EXPORTER_VERSION}/node_exporter-${NODE_EXPORTER_VERSION}.linux-arm64.tar.gz" \
+    "$NODE_EXPORTER_SHA256_ARM64" \
+    "node_exporter-${NODE_EXPORTER_VERSION}.linux-arm64/node_exporter" \
+    /usr/local/bin/prometheus-node-exporter
+
+  getent group prometheus >/dev/null || groupadd --system prometheus
+  id prometheus >/dev/null 2>&1 || useradd --system --gid prometheus --home-dir /var/lib/prometheus --shell /sbin/nologin prometheus
+  install -d -o root -g root -m 0755 /etc/prometheus
+  install -d -o prometheus -g prometheus -m 0750 /var/lib/prometheus/metrics2
+
+  cat >/etc/systemd/system/prometheus.service <<'EOF'
+[Unit]
+Description=Prometheus monitoring server
+Wants=network-online.target
+After=network-online.target
+
+[Service]
+Type=simple
+User=prometheus
+Group=prometheus
+Restart=on-failure
+RestartSec=5s
+ExecStart=/usr/local/bin/prometheus
+
+[Install]
+WantedBy=multi-user.target
+EOF
+  cat >/etc/systemd/system/prometheus-node-exporter.service <<'EOF'
+[Unit]
+Description=Prometheus Node Exporter
+Wants=network-online.target
+After=network-online.target
+
+[Service]
+Type=simple
+User=prometheus
+Group=prometheus
+Restart=on-failure
+RestartSec=5s
+ExecStart=/usr/local/bin/prometheus-node-exporter
+
+[Install]
+WantedBy=multi-user.target
+EOF
+}
+
+for command_name in systemctl install awk curl ss; do
+  command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }
+done
+
+if command -v apt-get >/dev/null 2>&1; then
+  log "Installing Prometheus and Node Exporter packages"
+  apt-get update
+  DEBIAN_FRONTEND=noninteractive apt-get install -y prometheus prometheus-node-exporter curl iproute2
+else
+  install_official_arm64_releases
+fi
 
 prometheus_bin="$(command -v prometheus)"
 node_exporter_bin="$(command -v prometheus-node-exporter)"
 promtool_bin="$(command -v promtool)"
 backup_dir="/etc/prometheus/cmpp-backups/$(date '+%Y%m%d-%H%M%S')"
-mkdir -p "$backup_dir" /etc/systemd/system/prometheus.service.d /etc/systemd/system/prometheus-node-exporter.service.d
+mkdir -p "$backup_dir" /etc/systemd/system/prometheus.service.d /etc/systemd/system/prometheus-node-exporter.service.d /etc/systemd/system/cmpp-api.service.d
 
 for config_file in /etc/prometheus/prometheus.yml /etc/prometheus/cmpp-alerts.yml /etc/prometheus/cmpp-alerts-source.yml; do
   if [[ -f "$config_file" ]]; then
@@ -37,6 +123,9 @@ fi
 if [[ -f /etc/systemd/system/prometheus-node-exporter.service.d/cmpp-monitoring.conf ]]; then
   cp --preserve=mode,timestamps /etc/systemd/system/prometheus-node-exporter.service.d/cmpp-monitoring.conf "$backup_dir/node-exporter-service-override.conf"
 fi
+if [[ -f /etc/systemd/system/cmpp-api.service.d/monitoring-storage.conf ]]; then
+  cp --preserve=mode,timestamps /etc/systemd/system/cmpp-api.service.d/monitoring-storage.conf "$backup_dir/cmpp-api-monitoring-storage.conf"
+fi
 
 log "Installing platform-owned scrape and alert configuration"
 install -o root -g root -m 0644 "$SCRIPT_DIR/prometheus.yml" /etc/prometheus/prometheus.yml
@@ -58,6 +147,14 @@ install -d -o cmpp-api -g prometheus -m 2750 /var/lib/cmpp-platform/monitoring
 if [[ ! -f /var/lib/cmpp-platform/monitoring/cmpp-managed-alerts.yml ]]; then
   install -o cmpp-api -g prometheus -m 0640 "$SCRIPT_DIR/cmpp-managed-alerts.yml" /var/lib/cmpp-platform/monitoring/cmpp-managed-alerts.yml
 fi
+chown cmpp-api:prometheus /var/lib/cmpp-platform/monitoring/cmpp-managed-alerts.yml
+chmod 0640 /var/lib/cmpp-platform/monitoring/cmpp-managed-alerts.yml
+
+# The API writes managed rules atomically. Its hardened unit otherwise sees /var/lib as read-only.
+cat >/etc/systemd/system/cmpp-api.service.d/monitoring-storage.conf <<'EOF'
+[Service]
+ReadWritePaths=/var/lib/cmpp-platform/monitoring
+EOF
 
 cat >/etc/systemd/system/prometheus.service.d/cmpp-monitoring.conf <<EOF
 [Service]
@@ -97,4 +194,4 @@ fi
 
 log "Prometheus monitoring is ready on loopback only"
 echo "Configuration backup: $backup_dir"
-echo "The CMPP API and Gateway were not restarted."
+echo "The CMPP API and Gateway were not restarted. Restart the API after setting PROMETHEUS_URL."

tools/monitoring/install-service-exporters.sh

@@ -1,10 +1,84 @@
 #!/usr/bin/env bash
 set -Eeuo pipefail
 
+POSTGRES_EXPORTER_VERSION="0.20.1"
+POSTGRES_EXPORTER_SHA256_ARM64="d5d86fb98bb1f26b088d1a6fda07fd6b6f035cb5d40492f75ec3bfebb5ddfe9d"
+REDIS_EXPORTER_VERSION="1.89.0"
+REDIS_EXPORTER_SHA256_ARM64="1a70d42cd1d96a9388f8b528f55ad5bdc2c3cbd911c1dff3875fcb9ac01fa28a"
+NGINX_EXPORTER_VERSION="1.5.3"
+NGINX_EXPORTER_SHA256_ARM64="a0b1a5f5bba09483bd2e04c759c1a75fffe46ca72c0314ee2bb925d1742ff23b"
+
 if [[ "$(id -u)" -ne 0 ]]; then echo "Run as root." >&2; exit 1; fi
-for command_name in apt-get systemctl nginx curl; do command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }; done
+for command_name in systemctl nginx curl install ss; do command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }; done
 
 log() { printf '\n[%s] %s\n' "$(date '+%F %T')" "$*"; }
+
+install_release_binary() {
+  local url="$1" expected_sha256="$2" archive_member="$3" destination="$4"
+  local work_dir archive
+  work_dir="$(mktemp -d /tmp/cmpp-monitoring.XXXXXX)"
+  archive="$work_dir/release.tar.gz"
+  curl -fL --retry 5 --retry-delay 2 --connect-timeout 15 -o "$archive" "$url"
+  printf '%s  %s\n' "$expected_sha256" "$archive" | sha256sum -c -
+  tar -xzf "$archive" -C "$work_dir"
+  install -o root -g root -m 0755 "$work_dir/$archive_member" "$destination"
+  rm -rf -- "$work_dir"
+}
+
+write_base_unit() {
+  local unit="$1" description="$2" executable="$3"
+  cat >"/etc/systemd/system/${unit}.service" <<EOF
+[Unit]
+Description=${description}
+Wants=network-online.target
+After=network-online.target
+
+[Service]
+Type=simple
+User=prometheus
+Group=prometheus
+Restart=on-failure
+RestartSec=5s
+ExecStart=${executable}
+
+[Install]
+WantedBy=multi-user.target
+EOF
+}
+
+install_official_arm64_releases() {
+  [[ "$(uname -m)" == "aarch64" ]] || {
+    echo "Official-binary installation currently supports Linux aarch64 only." >&2
+    exit 1
+  }
+  for command_name in sha256sum tar; do
+    command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }
+  done
+  getent group prometheus >/dev/null || groupadd --system prometheus
+  id prometheus >/dev/null 2>&1 || useradd --system --gid prometheus --home-dir /var/lib/prometheus --shell /sbin/nologin prometheus
+
+  log "Installing checksum-pinned official service exporter ARM64 releases"
+  install_release_binary \
+    "https://github.com/prometheus-community/postgres_exporter/releases/download/v${POSTGRES_EXPORTER_VERSION}/postgres_exporter-${POSTGRES_EXPORTER_VERSION}.linux-arm64.tar.gz" \
+    "$POSTGRES_EXPORTER_SHA256_ARM64" \
+    "postgres_exporter-${POSTGRES_EXPORTER_VERSION}.linux-arm64/postgres_exporter" \
+    /usr/local/bin/prometheus-postgres-exporter
+  install_release_binary \
+    "https://github.com/oliver006/redis_exporter/releases/download/v${REDIS_EXPORTER_VERSION}/redis_exporter-v${REDIS_EXPORTER_VERSION}.linux-arm64.tar.gz" \
+    "$REDIS_EXPORTER_SHA256_ARM64" \
+    "redis_exporter-v${REDIS_EXPORTER_VERSION}.linux-arm64/redis_exporter" \
+    /usr/local/bin/prometheus-redis-exporter
+  install_release_binary \
+    "https://github.com/nginx/nginx-prometheus-exporter/releases/download/v${NGINX_EXPORTER_VERSION}/nginx-prometheus-exporter_${NGINX_EXPORTER_VERSION}_linux_arm64.tar.gz" \
+    "$NGINX_EXPORTER_SHA256_ARM64" \
+    nginx-prometheus-exporter \
+    /usr/local/bin/prometheus-nginx-exporter
+
+  write_base_unit prometheus-postgres-exporter "Prometheus PostgreSQL Exporter" /usr/local/bin/prometheus-postgres-exporter
+  write_base_unit prometheus-redis-exporter "Prometheus Redis Exporter" /usr/local/bin/prometheus-redis-exporter
+  write_base_unit prometheus-nginx-exporter "Prometheus Nginx Exporter" /usr/local/bin/prometheus-nginx-exporter
+}
+
 backup_dir="/etc/prometheus/cmpp-backups/$(date '+%Y%m%d-%H%M%S')-service-exporters"
 mkdir -p "$backup_dir"
 
@@ -12,9 +86,13 @@ for config_file in /etc/cmpp-platform/minio.env /etc/nginx/conf.d/cmpp-monitorin
   [[ -f "$config_file" ]] && cp --preserve=mode,timestamps "$config_file" "$backup_dir/$(basename "$config_file")"
 done
 
-log "Installing bounded service exporters"
+if command -v apt-get >/dev/null 2>&1; then
-apt-get update
+  log "Installing bounded service exporter packages"
-DEBIAN_FRONTEND=noninteractive apt-get install -y prometheus-postgres-exporter prometheus-redis-exporter prometheus-nginx-exporter
+  apt-get update
+  DEBIAN_FRONTEND=noninteractive apt-get install -y prometheus-postgres-exporter prometheus-redis-exporter prometheus-nginx-exporter
+else
+  install_official_arm64_releases
+fi
 
 [[ -f /etc/cmpp-platform/cmpp-platform.env ]] || { echo "Missing platform environment file." >&2; exit 1; }
 set -a
@@ -26,26 +104,29 @@ cat >/etc/cmpp-platform/monitoring-exporters.env <<EOF
 DATA_SOURCE_NAME=${database_base}?sslmode=disable
 REDIS_ADDR=redis://${REDIS_HOST:-127.0.0.1}:${REDIS_PORT:-6379}
 EOF
-chmod 0600 /etc/cmpp-platform/monitoring-exporters.env
+chown root:prometheus /etc/cmpp-platform/monitoring-exporters.env
+chmod 0640 /etc/cmpp-platform/monitoring-exporters.env
 
 write_override() {
-  local unit="$1" executable="$2" arguments="$3"
+  local unit="$1" executable="$2" arguments="$3" environment_file="${4:-}"
   local directory="/etc/systemd/system/${unit}.service.d"
   mkdir -p "$directory"
   [[ -f "$directory/cmpp-monitoring.conf" ]] && cp --preserve=mode,timestamps "$directory/cmpp-monitoring.conf" "$backup_dir/${unit}-override.conf"
   cat >"$directory/cmpp-monitoring.conf" <<EOF
 [Service]
-EnvironmentFile=/etc/cmpp-platform/monitoring-exporters.env
 ExecStart=
 ExecStart=${executable} ${arguments}
 EOF
+  if [[ -n "$environment_file" ]]; then
+    sed -i "/^\[Service\]$/a EnvironmentFile=${environment_file}" "$directory/cmpp-monitoring.conf"
+  fi
 }
 
 postgres_exporter="$(command -v prometheus-postgres-exporter)"
 redis_exporter="$(command -v prometheus-redis-exporter)"
 nginx_exporter="$(command -v prometheus-nginx-exporter)"
-write_override prometheus-postgres-exporter "$postgres_exporter" '--web.listen-address=127.0.0.1:9187'
+write_override prometheus-postgres-exporter "$postgres_exporter" '--web.listen-address=127.0.0.1:9187' /etc/cmpp-platform/monitoring-exporters.env
-write_override prometheus-redis-exporter "$redis_exporter" '--web.listen-address=127.0.0.1:9121'
+write_override prometheus-redis-exporter "$redis_exporter" '--web.listen-address=127.0.0.1:9121' /etc/cmpp-platform/monitoring-exporters.env
 
 cat >/etc/nginx/conf.d/cmpp-monitoring-status.conf <<'EOF'
 server {