feat(watch): 监测槽 SSH 一键安装 psutil

探针报 psutil missing 时显示安装按钮；POST install-psutil 经 SSH pip/apt 安装并审计。
2026-06-12 04:04:48 +08:00
parent 32a1885f0d
commit 75efd506f5
12 changed files with 360 additions and 8 deletions
@@ -0,0 +1,36 @@
+# 审计 — 监测槽 SSH 安装 psutil
+
+**日期**：2026-06-12  
+**Changelog**: `docs/changelog/2026-06-12-watch-install-psutil-ssh.md`
+
+## 范围
+
+| 文件 |
+|------|
+| `server/utils/psutil_install.py` |
+| `server/application/services/watch_service.py` |
+| `server/api/watch.py` |
+| `frontend/src/components/watch/WatchSlotCard.vue` |
+| `frontend/src/components/watch/WatchSlotRow.vue` |
+| `frontend/src/utils/watchFormat.ts` |
+| `server/utils/watch_probe_errors.py` |
+| `tests/test_psutil_install.py` |
+| `tests/test_watch_pins.py` |
+
+## Step 3
+
+| H | 结论 |
+|---|------|
+| H1 鉴权 | PASS — 须 JWT；仅已 Pin 的服务器可安装 |
+| H2 吞错 | PASS — SSH/安装失败 422 + snackbar |
+| H4 审计 | PASS — `watch_install_psutil` |
+
+## Closure
+
+H1–H4 PASS
+
+## DoD
+
+- [x] pytest test_psutil_install + test_watch_pins
+- [x] npm run type-check
+- [ ] 生产浏览器：psutil 错误时出现按钮且安装后探针恢复
@@ -0,0 +1,40 @@
+# Changelog — 监测槽一键 SSH 安装 psutil
+
+**日期**：2026-06-12
+
+## 摘要
+
+监测槽探针报「子机未安装 psutil」时，卡片内显示 **「安装 psutil（SSH）」** 按钮，通过 SSH 在子机系统 Python 安装 psutil（pip --user → apt/yum → pip）。
+
+## 动机
+
+SSH 监测探针依赖系统 `python3` 的 psutil；Agent venv 内的 psutil 无法被探针脚本使用。手动 SSH 安装成本高。
+
+## 变更
+
+- 后端：`POST /api/watch/servers/{server_id}/install-psutil`（须已 Pin 该服务器）
+- `server/utils/psutil_install.py`：远程安装脚本 + 验证
+- 前端：`WatchSlotCard` 在 psutil 错误时显示按钮；`WatchSlotRow` 调用 API 并 refresh
+- 审计：`watch_install_psutil`
+
+## 涉及文件
+
+- `server/utils/psutil_install.py`（新）
+- `server/application/services/watch_service.py`
+- `server/api/watch.py`
+- `frontend/src/components/watch/WatchSlotCard.vue`
+- `frontend/src/components/watch/WatchSlotRow.vue`
+- `frontend/src/utils/watchFormat.ts`
+- `tests/test_psutil_install.py`（新）
+- `tests/test_watch_pins.py`
+
+## 迁移 / 重启
+
+- 无需迁移；需重启 API + 前端构建
+
+## 验证
+
+```bash
+.venv/bin/pytest tests/test_psutil_install.py tests/test_watch_pins.py -q
+cd frontend && npm run type-check
+```
@@ -1,11 +1,12 @@
 <script setup lang="ts">
 import { ref, computed } from 'vue'
 import type { WatchSlot } from '@/composables/useWatchPins'
-import { formatBytesPerSec, formatRemaining, probeStatusColor, probeStatusLabel, formatProbeError } from '@/utils/watchFormat'
+import { formatBytesPerSec, formatRemaining, probeStatusColor, probeStatusLabel, formatProbeError, isPsutilMissingError } from '@/utils/watchFormat'
 import WatchSparkline from '@/components/watch/WatchSparkline.vue'

 const props = defineProps<{
  slot: WatchSlot
+  psutilInstallLoading?: boolean
 }>()

 const emit = defineEmits<{
@@ -15,10 +16,14 @@ const emit = defineEmits<{
  pick: [slotIndex: number]
  monitoring: [slotIndex: number, enabled: boolean]
  ttl: [slotIndex: number, hours: 8 | 24]
+  installPsutil: [slotIndex: number]
 }>()

 const monitoringOn = computed(() => props.slot.monitoring_enabled !== false)
 const slotTtl = computed(() => (props.slot.ttl_hours === 24 ? 24 : 8) as 8 | 24)
+const showPsutilInstall = computed(() =>
+  monitoringOn.value && isPsutilMissingError(props.slot.metrics?.error, props.slot.metrics?.probe_status),
+)

 const chartMetric = ref<'cpu_pct' | 'mem_pct' | 'disk_pct'>('cpu_pct')

@@ -119,10 +124,23 @@ function pct(v: number | null | undefined) {

    <div
      v-if="slot.metrics?.error"
-      class="text-caption text-error mt-1 text-truncate"
+      class="text-caption text-error mt-1"
      :title="formatProbeError(slot.metrics.error, slot.metrics.probe_status)"
    >
-      {{ formatProbeError(slot.metrics.error, slot.metrics.probe_status) }}
+      <div class="text-truncate">
+        {{ formatProbeError(slot.metrics.error, slot.metrics.probe_status) }}
+      </div>
+      <v-btn
+        v-if="showPsutilInstall"
+        size="x-small"
+        variant="tonal"
+        color="primary"
+        class="mt-1"
+        :loading="props.psutilInstallLoading"
+        @click.stop="emit('installPsutil', slot.slot_index)"
+      >
+        安装 psutil（SSH）
+      </v-btn>
    </div>

    <div class="d-flex align-center justify-space-between mt-2">
@@ -1,5 +1,5 @@
 <script setup lang="ts">
-import { onMounted, ref } from 'vue'
+import { ref, onMounted } from 'vue'
 import { useRouter } from 'vue-router'
 import WatchSlotCard from '@/components/watch/WatchSlotCard.vue'
 import WatchProcessDrawer from '@/components/watch/WatchProcessDrawer.vue'
@@ -7,6 +7,7 @@ import WatchSlotPickDialog from '@/components/watch/WatchSlotPickDialog.vue'
 import { useWatchPins, type WatchSlot } from '@/composables/useWatchPins'
 import { useSnackbar } from '@/composables/useSnackbar'
 import { formatApiError } from '@/utils/apiError'
+import { http } from '@/api'

 const router = useRouter()
 const snackbar = useSnackbar()
@@ -17,6 +18,7 @@ const processSlot = ref<WatchSlot | null>(null)
 const showPickDialog = ref(false)
 const pickSlotIndex = ref<number | null>(null)
 const pickLoading = ref(false)
+const psutilLoadingSlot = ref<number | null>(null)

 onMounted(() => {
  refreshPins()
@@ -71,6 +73,24 @@ function onPickSlot(slotIndex: number) {
  showPickDialog.value = true
 }

+async function onInstallPsutil(slotIndex: number) {
+  const slot = slots.value.find((s) => s.slot_index === slotIndex)
+  if (!slot?.server_id) return
+  psutilLoadingSlot.value = slotIndex
+  try {
+    const data = await http.post<{ message?: string }>(
+      `/watch/servers/${slot.server_id}/install-psutil`,
+      {},
+    )
+    snackbar(data.message || 'psutil 安装成功', 'success')
+    await refreshPins()
+  } catch (e: unknown) {
+    snackbar(formatApiError(e, '安装 psutil 失败'), 'error')
+  } finally {
+    psutilLoadingSlot.value = null
+  }
+}
+
 async function onPickServer(serverId: number) {
  if (pickSlotIndex.value == null) return
  pickLoading.value = true
@@ -105,12 +125,14 @@ async function onPickServer(serverId: number) {
      <v-col v-for="slot in slots" :key="slot.slot_index" cols="12" sm="6" lg="3">
        <WatchSlotCard
          :slot="slot"
+          :psutil-install-loading="psutilLoadingSlot === slot.slot_index"
          @remove="onRemove"
          @processes="openProcesses"
          @history="openHistory"
          @pick="onPickSlot"
          @monitoring="onMonitoring"
          @ttl="onTtl"
+          @install-psutil="onInstallPsutil"
        />
      </v-col>
    </v-row>
@@ -27,7 +27,7 @@ const PROBE_STATUS_LABELS: Record<string, string> = {

 const PROBE_ERROR_EXACT: Record<string, string> = {
  parse_error: '探针返回数据无法解析',
-  'psutil missing': '子机未安装 psutil（请执行 pip install psutil）',
+  'psutil missing': '子机未安装 psutil',
  ssh_probe_failed: 'SSH 探针脚本执行失败',
  'SSH watch probe failed': 'SSH 监测探针失败',
  'SSH probe failed': 'SSH 探针失败',
@@ -94,6 +94,17 @@ export function formatProbeError(
  return `探针异常：${text}`
 }

+/** True when probe error indicates missing psutil on the remote host. */
+export function isPsutilMissingError(
+  error: string | null | undefined,
+  status?: string | null,
+): boolean {
+  const text = (error || '').toLowerCase()
+  if (text.includes('psutil missing')) return true
+  if (status?.toLowerCase().includes('parse_error') && text.includes('psutil')) return true
+  return false
+}
+
 export function probeStatusColor(status: string | null | undefined): string {
  if (!status || status === 'ok') return 'success'
  if (status.includes('timeout') || status.includes('auth')) return 'warning'
@@ -211,6 +211,25 @@ async def get_watch_processes(
    return await svc.get_processes(server_id)


+@router.post("/servers/{server_id}/install-psutil")
+async def install_watch_psutil(
+    server_id: int,
+    request: Request,
+    admin: Admin = Depends(get_current_admin),
+    db: AsyncSession = Depends(get_db),
+):
+    """Install psutil on the pinned server via SSH (for watch probe metrics)."""
+    svc = WatchService(db)
+    try:
+        return await svc.install_psutil_ssh(
+            admin=admin,
+            server_id=server_id,
+            ip_address=_client_ip(request),
+        )
+    except ValueError as e:
+        raise HTTPException(status_code=422, detail=str(e)) from e
+
+
 def _parse_server_ids(server_ids: str) -> list[int]:
    ids: list[int] = []
    for part in server_ids.split(","):
@@ -420,6 +420,45 @@ class WatchService:
        processes = await _load_processes(server_id)
        return {"server_id": server_id, "processes": processes or []}

+    async def install_psutil_ssh(
+        self,
+        *,
+        admin: Admin,
+        server_id: int,
+        ip_address: str = "",
+    ) -> dict[str, Any]:
+        pinned = await self.pinned_server_map(admin.id)
+        if server_id not in pinned:
+            raise ValueError("该服务器不在当前监测槽中")
+        server = await self.server_repo.get_by_id(server_id)
+        if not server:
+            raise ValueError("服务器不存在")
+
+        from server.utils.psutil_install import install_psutil_via_ssh
+
+        try:
+            result = await install_psutil_via_ssh(server)
+        except Exception as e:
+            raise ValueError(f"SSH 连接失败: {e}") from e
+
+        if not result.get("success"):
+            msg = str(result.get("message") or "psutil 安装失败")
+            raise ValueError(msg)
+
+        await self._audit(
+            admin=admin,
+            action="watch_install_psutil",
+            server=server,
+            detail=f"SSH 安装 psutil · {server.name or server_id}",
+            ip_address=ip_address,
+        )
+        await self.db.commit()
+        return {
+            "success": True,
+            "server_id": server_id,
+            "message": result.get("message") or "psutil 已就绪",
+        }
+
    async def _audit(
        self,
        *,
@@ -0,0 +1,81 @@
+"""Install psutil on managed servers via SSH (system Python for watch probe)."""
+
+from __future__ import annotations
+
+from server.application.server_batch_common import sudo_wrap
+from server.domain.models import Server
+from server.infrastructure.ssh.asyncssh_pool import exec_ssh_command
+
+# Try --user pip, then distro packages, then system pip. Verify with same python3 as watch probe.
+PSUTIL_INSTALL_SHELL = r"""
+set -e
+verify_psutil() {
+  for py in python3 python3.12 python3.11 python3.10; do
+    if command -v "$py" >/dev/null 2>&1 && "$py" -c "import psutil" 2>/dev/null; then
+      echo psutil_ok
+      exit 0
+    fi
+  done
+  return 1
+}
+verify_psutil || true
+
+for py in python3 python3.12 python3.11 python3.10; do
+  command -v "$py" >/dev/null 2>&1 || continue
+  if "$py" -m pip install --user -q psutil 2>/dev/null; then
+    verify_psutil && exit 0
+  fi
+done
+
+if command -v apt-get >/dev/null 2>&1; then
+  apt-get update -qq && DEBIAN_FRONTEND=noninteractive apt-get install -y -qq python3-psutil \
+    && verify_psutil && exit 0
+fi
+if command -v dnf >/dev/null 2>&1; then
+  dnf install -y -q python3-psutil && verify_psutil && exit 0
+fi
+if command -v yum >/dev/null 2>&1; then
+  yum install -y -q python3-psutil && verify_psutil && exit 0
+fi
+
+for py in python3 python3.12 python3.11 python3.10; do
+  command -v "$py" >/dev/null 2>&1 || continue
+  if "$py" -m pip install -q psutil 2>/dev/null; then
+    verify_psutil && exit 0
+  fi
+done
+
+echo "ERROR: psutil install failed — pip/apt unavailable or permission denied" >&2
+exit 1
+"""
+
+
+def build_psutil_install_cmd(*, ssh_user: str) -> str:
+    return sudo_wrap(PSUTIL_INSTALL_SHELL.strip(), ssh_user)
+
+
+async def install_psutil_via_ssh(server: Server, *, timeout: int = 90) -> dict[str, str | int]:
+    """Run remote install; returns {exit_code, stdout, stderr, already_installed?}."""
+    ssh_user = (server.username or "root").strip()
+    cmd = build_psutil_install_cmd(ssh_user=ssh_user)
+    result = await exec_ssh_command(server, cmd, timeout=timeout)
+    stdout = (result.get("stdout") or "").strip()
+    stderr = (result.get("stderr") or "").strip()
+    raw_exit = result.get("exit_code")
+    exit_code = int(raw_exit) if raw_exit is not None else 1
+    if exit_code == 0 and "psutil_ok" in stdout:
+        return {
+            "success": True,
+            "exit_code": 0,
+            "stdout": stdout[:2000],
+            "stderr": stderr[:500],
+            "message": "psutil 已就绪",
+        }
+    detail = stderr or stdout or f"exit {exit_code}"
+    return {
+        "success": False,
+        "exit_code": exit_code,
+        "stdout": stdout[:2000],
+        "stderr": stderr[:500],
+        "message": detail[:300],
+    }
@@ -4,7 +4,7 @@ from __future__ import annotations

 _EXACT: dict[str, str] = {
    "parse_error": "探针返回数据无法解析",
-    "psutil missing": "子机未安装 psutil（请执行 pip install psutil）",
+    "psutil missing": "子机未安装 psutil",
    "ssh_probe_failed": "SSH 探针脚本执行失败",
    "SSH watch probe failed": "SSH 监测探针失败",
    "SSH probe failed": "SSH 探针失败",
@@ -50,7 +50,7 @@ def watch_probe_error_zh(error: str | None, *, status: str | None = None) -> str
    if "permission denied" in lower or "authentication failed" in lower:
        return f"SSH 认证失败：{text}"
    if "psutil missing" in lower:
-        return "子机未安装 psutil（请执行 pip install psutil）"
+        return "子机未安装 psutil"
    if "ssh_probe_failed" in lower or "ssh watch probe failed" in lower:
        return "SSH 探针脚本执行失败"

@@ -0,0 +1,46 @@
+"""Tests for SSH psutil install helper."""
+
+from unittest.mock import AsyncMock, patch
+
+import pytest
+
+from server.utils.psutil_install import build_psutil_install_cmd, install_psutil_via_ssh
+
+
+def test_build_psutil_install_cmd_root():
+    cmd = build_psutil_install_cmd(ssh_user="root")
+    assert "pip install --user -q psutil" in cmd
+    assert "python3-psutil" in cmd
+    assert "verify_psutil" in cmd
+
+
+def test_build_psutil_install_cmd_non_root_wraps_sudo():
+    cmd = build_psutil_install_cmd(ssh_user="deploy")
+    assert "sudoers.d/nexus-agent" in cmd
+    assert "pip install --user -q psutil" in cmd
+
+
+@pytest.mark.asyncio
+async def test_install_psutil_via_ssh_success():
+    server = type("S", (), {"username": "root", "domain": "h.example.com", "port": 22})()
+    with patch(
+        "server.utils.psutil_install.exec_ssh_command",
+        new_callable=AsyncMock,
+        return_value={"exit_code": 0, "stdout": "psutil_ok\n", "stderr": ""},
+    ):
+        result = await install_psutil_via_ssh(server)
+    assert result["success"] is True
+    assert result["message"] == "psutil 已就绪"
+
+
+@pytest.mark.asyncio
+async def test_install_psutil_via_ssh_failure():
+    server = type("S", (), {"username": "root", "domain": "h.example.com", "port": 22})()
+    with patch(
+        "server.utils.psutil_install.exec_ssh_command",
+        new_callable=AsyncMock,
+        return_value={"exit_code": 1, "stdout": "", "stderr": "permission denied"},
+    ):
+        result = await install_psutil_via_ssh(server)
+    assert result["success"] is False
+    assert "permission denied" in result["message"]
@@ -206,6 +206,46 @@ async def test_watch_pin_ttl_expiry_pin_server_no_duplicate(db_session, test_adm
    assert result["slots"][0]["monitoring_enabled"] is True


+@pytest.mark.asyncio
+async def test_watch_install_psutil_requires_pin(db_session, test_admin, mock_watch_redis):
+    from server.application.services.watch_service import WatchService
+
+    server = Server(name="watch-psutil", domain="wp.example.com", port=22, agent_version="1.0")
+    db_session.add(server)
+    await db_session.commit()
+
+    svc = WatchService(db_session)
+    admin = test_admin["admin"]
+
+    with pytest.raises(ValueError, match="不在当前监测槽"):
+        await svc.install_psutil_ssh(admin=admin, server_id=server.id)
+
+
+@pytest.mark.asyncio
+async def test_watch_install_psutil_success(db_session, test_admin, mock_watch_redis, monkeypatch):
+    from server.application.services.watch_service import WatchService
+
+    server = Server(name="watch-psutil2", domain="wp2.example.com", port=22, agent_version="1.0")
+    db_session.add(server)
+    await db_session.commit()
+
+    svc = WatchService(db_session)
+    admin = test_admin["admin"]
+    await svc.pin_server(admin=admin, server_id=server.id)
+
+    async def _fake_install(_server):
+        return {"success": True, "message": "psutil 已就绪"}
+
+    monkeypatch.setattr(
+        "server.utils.psutil_install.install_psutil_via_ssh",
+        _fake_install,
+    )
+
+    result = await svc.install_psutil_ssh(admin=admin, server_id=server.id)
+    assert result["success"] is True
+    assert result["message"] == "psutil 已就绪"
+
+
@pytest.mark.asyncio
 async def test_watch_pin_ttl_expiry_pauses_not_deletes(db_session, test_admin, mock_watch_redis):
    from datetime import timedelta
@@ -5,7 +5,7 @@ from server.utils.watch_probe_errors import watch_probe_error_zh

 def test_parse_error_zh():
    assert watch_probe_error_zh("parse_error") == "探针返回数据无法解析"
-    assert watch_probe_error_zh("psutil missing") == "子机未安装 psutil（请执行 pip install psutil）"
+    assert watch_probe_error_zh("psutil missing") == "子机未安装 psutil"


 def test_parse_error_with_detail():