fix: correct operational statistics and form interactions

2026-09-09 23:15:42 +08:00
parent 6d63eb5452
commit 5bcdbb2a03
33 changed files with 2920 additions and 829 deletions
@@ -0,0 +1,45 @@
+import { alertHistoryRange, mergeAlertHistory } from './alert-history';
+
+describe('historical alert observation cycles', () => {
+  it('uses seven Shanghai calendar days and rejects invalid or excessive dates', () => {
+    expect(alertHistoryRange(undefined, undefined, new Date('2026-09-09T16:30:00Z'))).toMatchObject({
+      startDate: '2026-09-04',
+      endDate: '2026-09-10',
+    });
+    for (const [from, to] of [
+      ['2026-02-30', '2026-03-01'],
+      ['2026-09-09', '2026-09-08'],
+      ['2026-07-01', '2026-09-09'],
+    ]) {
+      expect(() => alertHistoryRange(from, to)).toThrow();
+    }
+  });
+  it('retains distinct cycles, merges daily boundaries and excludes stale/nonpositive samples', () => {
+    const result = new Map();
+    const metric = { __name__: 'ALERTS_FOR_STATE', alertname: 'CPUHigh', instance: 'host', severity: 'warning' };
+    mergeAlertHistory(
+      result,
+      [
+        {
+          metric,
+          values: [
+            [110, '100'],
+            [120, '100'],
+            [130, '0'],
+            [140, 'NaN'],
+            [150, '145'],
+            [200, '145'],
+          ],
+        },
+      ],
+      110,
+      200,
+    );
+    mergeAlertHistory(result, [{ metric, values: [[160, '145']] }], 110, 200);
+    expect(result.size).toBe(2);
+    expect([...result.values()].map((item) => item.lastObservedAt)).toEqual([
+      new Date(120000).toISOString(),
+      new Date(160000).toISOString(),
+    ]);
+  });
+});
@@ -0,0 +1,68 @@
+import { BadRequestException } from '@nestjs/common';
+import { createHash } from 'node:crypto';
+
+export function alertHistoryRange(from?: string, to?: string, now = new Date()) {
+  const dateKey = (date: Date) => new Date(date.getTime() + 8 * 3600_000).toISOString().slice(0, 10);
+  const endDate = to || dateKey(now);
+  const startDate = from || dateKey(new Date(now.getTime() - 6 * 86400_000));
+  const parse = (value: string) => {
+    const result = new Date(`${value}T00:00:00+08:00`);
+    if (!/^\d{4}-\d{2}-\d{2}$/.test(value) || !Number.isFinite(result.getTime()) || dateKey(result) !== value) {
+      throw new BadRequestException('告警日期无效');
+    }
+    return result.getTime() / 1000;
+  };
+  const start = parse(startDate);
+  const end = parse(endDate) + 86400;
+  if (end <= start || end - start > 31 * 86400) throw new BadRequestException('告警日期范围须为1至31天');
+  return { startDate, endDate, start, end: Math.min(end, now.getTime() / 1000) };
+}
+
+export type AlertHistoryItem = {
+  id: string;
+  name: string;
+  severity: string;
+  service: string;
+  instance: string;
+  startedAt: string;
+  firstObservedAt: string;
+  lastObservedAt: string;
+};
+
+// ALERTS_FOR_STATE stores activeAt as the sample value, separating repeated trigger cycles.
+// Observation boundaries are not claimed as exact recovery times.
+export function mergeAlertHistory(
+  target: Map<string, AlertHistoryItem>,
+  series: Array<{ metric: Record<string, string>; values?: [number, string][] }>,
+  start: number,
+  end: number,
+) {
+  for (const { metric, values } of series) {
+    const labels = Object.entries(metric)
+      .filter(([key]) => key !== '__name__')
+      .sort(([a], [b]) => a.localeCompare(b));
+    const fingerprint = createHash('sha256').update(JSON.stringify(labels)).digest('hex');
+    for (const [time, rawActiveAt] of values ?? []) {
+      const activeAt = Number(rawActiveAt);
+      if (time < start || time >= end || !Number.isFinite(activeAt) || activeAt <= 0 || activeAt > time) continue;
+      const id = `${fingerprint}:${activeAt}`;
+      const observed = new Date(time * 1000).toISOString();
+      const item = target.get(id);
+      if (item) {
+        if (observed < item.firstObservedAt) item.firstObservedAt = observed;
+        if (observed > item.lastObservedAt) item.lastObservedAt = observed;
+      } else {
+        target.set(id, {
+          id,
+          name: metric.alertname || '未命名告警',
+          severity: metric.severity || 'info',
+          service: metric.service || '',
+          instance: metric.instance || '',
+          startedAt: new Date(activeAt * 1000).toISOString(),
+          firstObservedAt: observed,
+          lastObservedAt: observed,
+        });
+      }
+    }
+  }
+}
@@ -8,7 +8,10 @@ import { InfrastructureMonitoringService } from './infrastructure-monitoring.ser
@ApiTags('infrastructure-monitoring')
@Controller('admin/infrastructure-monitoring')
 export class InfrastructureMonitoringController {
-  constructor(private readonly monitoring: InfrastructureMonitoringService, private readonly settings: InfrastructureAlertSettingsService) {}
+  constructor(
+    private readonly monitoring: InfrastructureMonitoringService,
+    private readonly settings: InfrastructureAlertSettingsService,
+  ) {}

  @Get('overview')
  overview(@Query('range') range?: string, @CurrentSessionUserId() userId?: string) {
@@ -16,19 +19,35 @@ export class InfrastructureMonitoringController {
  }

  @Get('notification-summary')
-  notificationSummary(@CurrentSessionUserId() userId?: string) { return this.monitoring.notificationSummary(userId); }
+  notificationSummary(@CurrentSessionUserId() userId?: string) {
+    return this.monitoring.notificationSummary(userId);
+  }
+
+  @Get('alert-history')
+  alertHistory(@Query('from') from?: string, @Query('to') to?: string, @Query('page') page?: string) {
+    return this.monitoring.alertHistory(from, to, page);
+  }

  @Post('alerts/:fingerprint/read')
-  markAlertRead(@Param('fingerprint') fingerprint: string, @Body('activeAt') activeAt: unknown, @CurrentSessionUserId() userId: string) {
+  markAlertRead(
+    @Param('fingerprint') fingerprint: string,
+    @Body('activeAt') activeAt: unknown,
+    @CurrentSessionUserId() userId: string,
+  ) {
    return this.monitoring.markAlertRead(fingerprint, activeAt, userId);
  }

  @Get('alert-thresholds')
-  alertThresholds() { return this.settings.get(); }
+  alertThresholds() {
+    return this.settings.get();
+  }

  @Put('alert-thresholds')
  @RequireRecentAuthentication()
-  updateAlertThresholds(@Body() body: { configVersion?: number; thresholds?: unknown }, @CurrentSessionUserId() operatorId?: string) {
+  updateAlertThresholds(
+    @Body() body: { configVersion?: number; thresholds?: unknown },
+    @CurrentSessionUserId() operatorId?: string,
+  ) {
    return this.settings.update(body, operatorId);
  }
 }
@@ -1,9 +1,22 @@
-import { BadRequestException, Injectable, Logger, NotFoundException, ServiceUnavailableException } from '@nestjs/common';
+import {
+  BadRequestException,
+  Injectable,
+  Logger,
+  NotFoundException,
+  ServiceUnavailableException,
+} from '@nestjs/common';
 import { ConfigService } from '@nestjs/config';
 import { Prisma } from '@prisma/client';
 import { createHash } from 'node:crypto';
 import { PrismaService } from '../prisma/prisma.service';
-import { compareMountpoints, FILESYSTEM_LABELS, FILESYSTEM_SELECTOR, FILESYSTEM_USAGE_PERCENT, filesystemIdentity } from './filesystem-metrics';
+import { alertHistoryRange, mergeAlertHistory, type AlertHistoryItem } from './alert-history';
+import {
+  compareMountpoints,
+  FILESYSTEM_LABELS,
+  FILESYSTEM_SELECTOR,
+  FILESYSTEM_USAGE_PERCENT,
+  filesystemIdentity,
+} from './filesystem-metrics';
 import type {
  InfrastructureAlert,
  InfrastructureMetricPoint,
@@ -57,7 +70,8 @@ const QUERIES = {
  uptimeSeconds: 'time() - node_boot_time_seconds',
  lastSampleAt: 'max(timestamp(node_uname_info))',
  // PromQL字符串本身需要两个反斜杠才能把正则的“\.”传给RE2；TypeScript字面量因此需要写四个。
-  services: 'max by (name) (node_systemd_unit_state{name=~"cmpp-api\\\\.service|cmpp-gateway\\\\.service|postgresql\\\\.service|redis(-server)?\\\\.service|cmpp-minio\\\\.service|nginx\\\\.service",state="active"})',
+  services:
+    'max by (name) (node_systemd_unit_state{name=~"cmpp-api\\\\.service|cmpp-gateway\\\\.service|postgresql\\\\.service|redis(-server)?\\\\.service|cmpp-minio\\\\.service|nginx\\\\.service",state="active"})',
 } as const;

 const SERVICE_DEFINITIONS = [
@@ -70,38 +84,62 @@ const SERVICE_DEFINITIONS = [
 ] as const;

 const SERVICE_METRIC_DEFINITIONS = [
-  { key: 'api', name: 'API服务', metrics: [
-    ['requestsPerSecond', '请求速率', 'cmpp:service_api:requests_per_second', 'per_second'],
-    ['errorPercent', '5xx错误率', 'cmpp:service_api:error_percent', 'percent'],
-    ['latencyP95', 'P95响应', 'cmpp:service_api:latency_p95_seconds', 'seconds'],
-    ['eventLoopP99', '事件循环P99', 'cmpp:service_api:event_loop_p99_seconds', 'seconds'],
-  ] },
-  { key: 'gateway', name: 'Gateway服务', metrics: [
-    ['submitsPerSecond', '提交速率', 'cmpp:service_gateway:submits_per_second', 'per_second'],
-    ['failurePercent', '提交失败率', 'cmpp:service_gateway:failure_percent', 'percent'],
-    ['queuePending', 'Stream pending', 'cmpp:service_gateway:queue_pending', 'count'],
-    ['queueOldestSeconds', '最旧pending', 'cmpp:service_gateway:queue_oldest_seconds', 'seconds'],
-  ] },
-  { key: 'postgresql', name: 'PostgreSQL', metrics: [
-    ['connectionPercent', '连接使用率', 'cmpp:service_postgresql:connection_percent', 'percent'],
-    ['deadlocks15m', '15分钟死锁', 'cmpp:service_postgresql:deadlocks_15m', 'count'],
-  ] },
-  { key: 'redis', name: 'Redis', metrics: [
-    ['memoryPercent', '内存使用率', 'cmpp:service_redis:memory_percent', 'percent'],
-    ['memoryUsedBytes', '已用内存', 'cmpp:service_redis:memory_used_bytes', 'bytes'],
-    ['connectedClients', '客户端连接', 'cmpp:service_redis:connected_clients', 'count'],
-    ['evictions5m', '5分钟淘汰', 'cmpp:service_redis:evictions_5m', 'count'],
-  ] },
-  { key: 'minio', name: 'MinIO', metrics: [
-    ['capacityPercent', '存储容量使用率', 'cmpp:service_minio:capacity_percent', 'percent'],
-    ['usageBytes', '对象数据量', 'cmpp:service_minio:usage_bytes', 'bytes'],
-    ['objects', '对象数', 'cmpp:service_minio:objects', 'count'],
-    ['drivesOffline', '离线存储盘', 'cmpp:service_minio:drives_offline', 'count'],
-  ] },
-  { key: 'nginx', name: 'Nginx', metrics: [
-    ['connectionsActive', '活跃连接', 'cmpp:service_nginx:connections_active', 'count'],
-    ['requestsPerSecond', '请求速率', 'cmpp:service_nginx:requests_per_second', 'per_second'],
-  ] },
+  {
+    key: 'api',
+    name: 'API服务',
+    metrics: [
+      ['requestsPerSecond', '请求速率', 'cmpp:service_api:requests_per_second', 'per_second'],
+      ['errorPercent', '5xx错误率', 'cmpp:service_api:error_percent', 'percent'],
+      ['latencyP95', 'P95响应', 'cmpp:service_api:latency_p95_seconds', 'seconds'],
+      ['eventLoopP99', '事件循环P99', 'cmpp:service_api:event_loop_p99_seconds', 'seconds'],
+    ],
+  },
+  {
+    key: 'gateway',
+    name: 'Gateway服务',
+    metrics: [
+      ['submitsPerSecond', '提交速率', 'cmpp:service_gateway:submits_per_second', 'per_second'],
+      ['failurePercent', '提交失败率', 'cmpp:service_gateway:failure_percent', 'percent'],
+      ['queuePending', 'Stream pending', 'cmpp:service_gateway:queue_pending', 'count'],
+      ['queueOldestSeconds', '最旧pending', 'cmpp:service_gateway:queue_oldest_seconds', 'seconds'],
+    ],
+  },
+  {
+    key: 'postgresql',
+    name: 'PostgreSQL',
+    metrics: [
+      ['connectionPercent', '连接使用率', 'cmpp:service_postgresql:connection_percent', 'percent'],
+      ['deadlocks15m', '15分钟死锁', 'cmpp:service_postgresql:deadlocks_15m', 'count'],
+    ],
+  },
+  {
+    key: 'redis',
+    name: 'Redis',
+    metrics: [
+      ['memoryPercent', '内存使用率', 'cmpp:service_redis:memory_percent', 'percent'],
+      ['memoryUsedBytes', '已用内存', 'cmpp:service_redis:memory_used_bytes', 'bytes'],
+      ['connectedClients', '客户端连接', 'cmpp:service_redis:connected_clients', 'count'],
+      ['evictions5m', '5分钟淘汰', 'cmpp:service_redis:evictions_5m', 'count'],
+    ],
+  },
+  {
+    key: 'minio',
+    name: 'MinIO',
+    metrics: [
+      ['capacityPercent', '存储容量使用率', 'cmpp:service_minio:capacity_percent', 'percent'],
+      ['usageBytes', '对象数据量', 'cmpp:service_minio:usage_bytes', 'bytes'],
+      ['objects', '对象数', 'cmpp:service_minio:objects', 'count'],
+      ['drivesOffline', '离线存储盘', 'cmpp:service_minio:drives_offline', 'count'],
+    ],
+  },
+  {
+    key: 'nginx',
+    name: 'Nginx',
+    metrics: [
+      ['connectionsActive', '活跃连接', 'cmpp:service_nginx:connections_active', 'count'],
+      ['requestsPerSecond', '请求速率', 'cmpp:service_nginx:requests_per_second', 'per_second'],
+    ],
+  },
 ] as const;

 const SERVICE_METRICS_QUERY = '{__name__=~"cmpp:service_.*"}';
@@ -165,7 +203,10 @@ export class InfrastructureMonitoringService {
  private readonly prometheusUrl: string;
  private readonly queryTimeoutMs: number;

-  constructor(config: ConfigService, private readonly prisma: PrismaService) {
+  constructor(
+    config: ConfigService,
+    private readonly prisma: PrismaService,
+  ) {
    this.prometheusUrl = normalizePrometheusUrl(config.get('PROMETHEUS_URL'));
    this.queryTimeoutMs = Math.min(15_000, Math.max(1_000, Number(config.get('PROMETHEUS_QUERY_TIMEOUT_MS') ?? 5_000)));
  }
@@ -202,7 +243,7 @@ export class InfrastructureMonitoringService {
          activeAlerts: alerts.length,
        },
        metrics: instant.metrics,
-        trends: { ...trends.metrics, diskUsagePercent: rootDisk ? trends.disks.get(rootDisk.id) ?? [] : [] },
+        trends: { ...trends.metrics, diskUsagePercent: rootDisk ? (trends.disks.get(rootDisk.id) ?? []) : [] },
        disks: instant.disks.map((disk) => ({ ...disk, trend: trends.disks.get(disk.id) ?? [] })),
        services,
        serviceMetrics,
@@ -210,18 +251,28 @@ export class InfrastructureMonitoringService {
      };
    } catch (error) {
      // 页面必须整体清空陈旧指标，但服务端仍需留下不含PromQL/地址/凭据的根因摘要便于运维诊断。
-      this.logger.warn(`Prometheus monitoring overview unavailable: ${error instanceof Error ? error.message : 'unknown error'}`);
+      this.logger.warn(
+        `Prometheus monitoring overview unavailable: ${error instanceof Error ? error.message : 'unknown error'}`,
+      );
      return this.unavailable(range, collectedAt);
    }
  }

  async notificationSummary(userId?: string) {
    try {
-      const alerts = await this.attachReadState(this.parseAlerts(await this.getJson<PrometheusAlertResponse>('/api/v1/alerts')), userId);
+      const alerts = await this.attachReadState(
+        this.parseAlerts(await this.getJson<PrometheusAlertResponse>('/api/v1/alerts')),
+        userId,
+      );
      const unreadAlerts = alerts.filter((item) => !item.acknowledged);
-      return { count: unreadAlerts.length, criticalCount: unreadAlerts.filter((item) => item.severity === 'critical').length };
+      return {
+        count: unreadAlerts.length,
+        criticalCount: unreadAlerts.filter((item) => item.severity === 'critical').length,
+      };
    } catch (error) {
-      this.logger.warn(`Prometheus notification summary unavailable: ${error instanceof Error ? error.message : 'unknown error'}`);
+      this.logger.warn(
+        `Prometheus notification summary unavailable: ${error instanceof Error ? error.message : 'unknown error'}`,
+      );
      throw new ServiceUnavailableException('Prometheus活动告警当前不可用');
    }
  }
@@ -231,12 +282,21 @@ export class InfrastructureMonitoringService {
    const activeAt = new Date(String(rawActiveAt ?? ''));
    if (!Number.isFinite(activeAt.getTime())) throw new BadRequestException('告警开始时间无效');
    const activeAlerts = this.parseAlerts(await this.getJson<PrometheusAlertResponse>('/api/v1/alerts'));
-    const current = activeAlerts.find((item) => item.fingerprint === fingerprint && Date.parse(item.startedAt) === activeAt.getTime());
+    const current = activeAlerts.find(
+      (item) => item.fingerprint === fingerprint && Date.parse(item.startedAt) === activeAt.getTime(),
+    );
    if (!current) throw new NotFoundException('该次活动告警已结束或已重新触发，请刷新后重试');
    const readAt = new Date();
-    const log = () => this.prisma.operationLog.create({
-      data: { userId, action: 'monitoring.alert_marked_read', resource: 'infrastructure_alert', resourceId: fingerprint, detail: { activeAt: activeAt.toISOString(), alertName: current.name, severity: current.severity } },
-    });
+    const log = () =>
+      this.prisma.operationLog.create({
+        data: {
+          userId,
+          action: 'monitoring.alert_marked_read',
+          resource: 'infrastructure_alert',
+          resourceId: fingerprint,
+          detail: { activeAt: activeAt.toISOString(), alertName: current.name, severity: current.severity },
+        },
+      });
    let read;
    try {
      [read] = await this.prisma.$transaction([
@@ -245,15 +305,57 @@ export class InfrastructureMonitoringService {
      ]);
    } catch (error) {
      if (!(error instanceof Prisma.PrismaClientKnownRequestError) || error.code !== 'P2002') throw error;
-      const existing = await this.prisma.infrastructureAlertRead.findUniqueOrThrow({ where: { fingerprint_userId: { fingerprint, userId } } });
+      const existing = await this.prisma.infrastructureAlertRead.findUniqueOrThrow({
+        where: { fingerprint_userId: { fingerprint, userId } },
+      });
      // 同一次触发重复点击不更新readAt也不重复写日志；activeAt变化才代表同指纹的新触发周期。
      if (existing.activeAt.getTime() === activeAt.getTime()) read = existing;
-      else [read] = await this.prisma.$transaction([
-        this.prisma.infrastructureAlertRead.update({ where: { fingerprint_userId: { fingerprint, userId } }, data: { activeAt, readAt } }),
-        log(),
-      ]);
+      else
+        [read] = await this.prisma.$transaction([
+          this.prisma.infrastructureAlertRead.update({
+            where: { fingerprint_userId: { fingerprint, userId } },
+            data: { activeAt, readAt },
+          }),
+          log(),
+        ]);
    }
-    return { fingerprint, activeAt: read.activeAt.toISOString(), acknowledged: true, acknowledgedAt: read.readAt.toISOString() };
+    return {
+      fingerprint,
+      activeAt: read.activeAt.toISOString(),
+      acknowledged: true,
+      acknowledgedAt: read.readAt.toISOString(),
+    };
+  }
+
+  async alertHistory(from?: string, to?: string, rawPage?: string) {
+    const range = alertHistoryRange(from, to);
+    const page = rawPage === undefined ? 1 : Number(rawPage);
+    if (!Number.isSafeInteger(page) || page < 1) throw new BadRequestException('告警页码无效');
+    const history = new Map<string, AlertHistoryItem>();
+    try {
+      // Daily raw range vectors retain short events that a coarse query_range step would miss.
+      for (let start = range.start; start < range.end; start += 86400) {
+        const end = Math.min(start + 86400, range.end);
+        const response = await this.getJson<PrometheusQueryResponse>('/api/v1/query', {
+          query: `ALERTS_FOR_STATE[${Math.ceil(end - start)}s]`,
+          time: String(end),
+        });
+        mergeAlertHistory(history, response.data?.result ?? [], range.start, range.end);
+      }
+    } catch {
+      throw new ServiceUnavailableException('历史告警查询失败，请稍后重试');
+    }
+    const items = [...history.values()].sort(
+      (a, b) => b.startedAt.localeCompare(a.startedAt) || a.id.localeCompare(b.id),
+    );
+    return {
+      items: items.slice((page - 1) * 25, page * 25),
+      total: items.length,
+      page,
+      pageSize: 25,
+      startDate: range.startDate,
+      endDate: range.endDate,
+    };
  }

  private parseRange(value?: string): InfrastructureMonitoringRange {
@@ -264,12 +366,21 @@ export class InfrastructureMonitoringService {

  private async loadInstantMetrics() {
    const keys = Object.keys(emptyMetrics()) as Array<keyof InfrastructureMonitoringOverview['metrics']>;
-    const responses = await Promise.all([...keys.map((key) => this.query(QUERIES[key])), this.query(QUERIES.lastSampleAt)]);
+    const responses = await Promise.all([
+      ...keys.map((key) => this.query(QUERIES[key])),
+      this.query(QUERIES.lastSampleAt),
+    ]);
    const metrics = emptyMetrics();
-    keys.forEach((key, index) => { if (!key.startsWith('disk')) metrics[key] = vectorValue(responses[index]); });
+    keys.forEach((key, index) => {
+      if (!key.startsWith('disk')) metrics[key] = vectorValue(responses[index]);
+    });
    const diskSamples = (key: keyof typeof metrics) => responses[keys.indexOf(key)].data?.result ?? [];
-    const usage = new Map(diskSamples('diskUsagePercent').map((item) => [filesystemIdentity(item.metric), finiteNumber(item.value?.[1])]));
-    const available = new Map(diskSamples('diskAvailableBytes').map((item) => [filesystemIdentity(item.metric), finiteNumber(item.value?.[1])]));
+    const usage = new Map(
+      diskSamples('diskUsagePercent').map((item) => [filesystemIdentity(item.metric), finiteNumber(item.value?.[1])]),
+    );
+    const available = new Map(
+      diskSamples('diskAvailableBytes').map((item) => [filesystemIdentity(item.metric), finiteNumber(item.value?.[1])]),
+    );
    const groups = new Map<string, PrometheusSeries[]>();
    for (const item of diskSamples('diskTotalBytes')) {
      if (!item.metric.device || !item.metric.mountpoint || (finiteNumber(item.value?.[1]) ?? 0) <= 0) continue;
@@ -278,18 +389,32 @@ export class InfrastructureMonitoringService {
      group.push(item);
      groups.set(id, group);
    }
-    const disks = [...groups].map(([id, items]) => {
-      const mountpoints = [...new Set(items.map((item) => item.metric.mountpoint))].sort(compareMountpoints);
-      const metric = items[0].metric;
-      return {
-        id, instance: metric.instance ?? '', device: metric.device, filesystem: metric.fstype ?? '',
-        mountpoint: mountpoints[0], mountpoints,
-        // Never sum aliases. Max/min also tolerate slight sampling differences.
-        totalBytes: Math.max(...items.map((item) => finiteNumber(item.value?.[1])!)),
-        availableBytes: available.get(id) ?? null, usagePercent: usage.get(id) ?? null,
-      };
-    })
-      .sort((left, right) => left.instance.localeCompare(right.instance) || (left.mountpoint === '/' ? -1 : right.mountpoint === '/' ? 1 : left.mountpoint.localeCompare(right.mountpoint)));
+    const disks = [...groups]
+      .map(([id, items]) => {
+        const mountpoints = [...new Set(items.map((item) => item.metric.mountpoint))].sort(compareMountpoints);
+        const metric = items[0].metric;
+        return {
+          id,
+          instance: metric.instance ?? '',
+          device: metric.device,
+          filesystem: metric.fstype ?? '',
+          mountpoint: mountpoints[0],
+          mountpoints,
+          // Never sum aliases. Max/min also tolerate slight sampling differences.
+          totalBytes: Math.max(...items.map((item) => finiteNumber(item.value?.[1])!)),
+          availableBytes: available.get(id) ?? null,
+          usagePercent: usage.get(id) ?? null,
+        };
+      })
+      .sort(
+        (left, right) =>
+          left.instance.localeCompare(right.instance) ||
+          (left.mountpoint === '/'
+            ? -1
+            : right.mountpoint === '/'
+              ? 1
+              : left.mountpoint.localeCompare(right.mountpoint)),
+      );
    const rootDisk = disks.find((disk) => disk.mountpoints.includes('/'));
    metrics.diskUsagePercent = rootDisk?.usagePercent ?? null;
    metrics.diskTotalBytes = rootDisk?.totalBytes ?? null;
@@ -304,21 +429,32 @@ export class InfrastructureMonitoringService {
    const keys = Object.keys(emptyTrends()) as Array<keyof InfrastructureMonitoringOverview['trends']>;
    const responses = await Promise.all(keys.map((key) => this.queryRange(QUERIES[key], start, end, config.step)));
    return {
-      metrics: Object.fromEntries(keys.map((key, index) => [key, key === 'diskUsagePercent' ? [] : matrixValues(responses[index])])) as InfrastructureMonitoringOverview['trends'],
-      disks: new Map((responses[keys.indexOf('diskUsagePercent')].data?.result ?? []).map((item) => [
-        filesystemIdentity(item.metric), matrixValues({ status: 'success', data: { result: [item] } }),
-      ])),
+      metrics: Object.fromEntries(
+        keys.map((key, index) => [key, key === 'diskUsagePercent' ? [] : matrixValues(responses[index])]),
+      ) as InfrastructureMonitoringOverview['trends'],
+      disks: new Map(
+        (responses[keys.indexOf('diskUsagePercent')].data?.result ?? []).map((item) => [
+          filesystemIdentity(item.metric),
+          matrixValues({ status: 'success', data: { result: [item] } }),
+        ]),
+      ),
    };
  }

  private parseServices(response: PrometheusQueryResponse): InfrastructureServiceStatus[] {
    const values = new Map<string, number>();
    for (const item of response.data?.result ?? []) {
-      if (item.metric.name) values.set(item.metric.name, vectorValue({ status: 'success', data: { result: [item] } }) ?? 0);
+      if (item.metric.name)
+        values.set(item.metric.name, vectorValue({ status: 'success', data: { result: [item] } }) ?? 0);
    }
    return SERVICE_DEFINITIONS.map((definition) => {
      const present = definition.units.filter((unit) => values.has(unit));
-      const status = present.length === 0 ? 'unknown' : present.some((unit) => (values.get(unit) ?? 0) >= 1) ? 'healthy' : 'unhealthy';
+      const status =
+        present.length === 0
+          ? 'unknown'
+          : present.some((unit) => (values.get(unit) ?? 0) >= 1)
+            ? 'healthy'
+            : 'unhealthy';
      return { key: definition.key, name: definition.name, unit: present[0] ?? definition.units[0], status };
    });
  }
@@ -329,7 +465,8 @@ export class InfrastructureMonitoringService {
      .map<InfrastructureAlert>((item) => {
        const labels = item.labels ?? {};
        const annotations = item.annotations ?? {};
-        const severity: InfrastructureAlert['severity'] = labels.severity === 'critical' ? 'critical' : labels.severity === 'warning' ? 'warning' : 'info';
+        const severity: InfrastructureAlert['severity'] =
+          labels.severity === 'critical' ? 'critical' : labels.severity === 'warning' ? 'warning' : 'info';
        const identity = JSON.stringify(Object.entries(labels).sort(([left], [right]) => left.localeCompare(right)));
        return {
          fingerprint: createHash('sha256').update(identity).digest('hex').slice(0, 24),
@@ -348,7 +485,9 @@ export class InfrastructureMonitoringService {
      })
      .sort((left, right) => {
        const priority: Record<InfrastructureAlert['severity'], number> = { critical: 0, warning: 1, info: 2 };
-        return priority[left.severity] - priority[right.severity] || Date.parse(left.startedAt) - Date.parse(right.startedAt);
+        return (
+          priority[left.severity] - priority[right.severity] || Date.parse(left.startedAt) - Date.parse(right.startedAt)
+        );
      });
  }

@@ -377,24 +516,46 @@ export class InfrastructureMonitoringService {
      key: group.key,
      name: group.name,
      available: group.metrics.some((metric) => values.has(metric[2])),
-      metrics: group.metrics.map(([key, label, metricName, unit]) => ({ key, label, value: values.get(metricName) ?? null, unit })),
+      metrics: group.metrics.map(([key, label, metricName, unit]) => ({
+        key,
+        label,
+        value: values.get(metricName) ?? null,
+        unit,
+      })),
    }));
  }

  private unavailable(range: InfrastructureMonitoringRange, collectedAt: string): InfrastructureMonitoringOverview {
-    const services = SERVICE_DEFINITIONS.map((item) => ({ key: item.key, name: item.name, unit: item.units[0], status: 'unknown' as const }));
+    const services = SERVICE_DEFINITIONS.map((item) => ({
+      key: item.key,
+      name: item.name,
+      unit: item.units[0],
+      status: 'unknown' as const,
+    }));
    return {
      available: false,
      range,
      collectedAt,
      lastSampleAt: null,
      error: 'Prometheus监控数据当前不可用，请检查采集与服务状态',
-      summary: { overallStatus: 'unknown', serviceTotal: services.length, serviceHealthy: 0, warningAlerts: 0, criticalAlerts: 0, activeAlerts: 0 },
+      summary: {
+        overallStatus: 'unknown',
+        serviceTotal: services.length,
+        serviceHealthy: 0,
+        warningAlerts: 0,
+        criticalAlerts: 0,
+        activeAlerts: 0,
+      },
      metrics: emptyMetrics(),
      disks: [],
      trends: emptyTrends(),
      services,
-      serviceMetrics: SERVICE_METRIC_DEFINITIONS.map((group) => ({ key: group.key, name: group.name, available: false, metrics: [] })),
+      serviceMetrics: SERVICE_METRIC_DEFINITIONS.map((group) => ({
+        key: group.key,
+        name: group.name,
+        available: false,
+        metrics: [],
+      })),
      alerts: [],
    };
  }
@@ -404,15 +565,26 @@ export class InfrastructureMonitoringService {
  }

  private queryRange(query: string, start: number, end: number, step: number) {
-    return this.getJson<PrometheusQueryResponse>('/api/v1/query_range', { query, start: String(start), end: String(end), step: String(step) });
+    return this.getJson<PrometheusQueryResponse>('/api/v1/query_range', {
+      query,
+      start: String(start),
+      end: String(end),
+      step: String(step),
+    });
  }

-  private async getJson<T extends { status: 'success' | 'error'; error?: string }>(path: string, params: Record<string, string> = {}): Promise<T> {
+  private async getJson<T extends { status: 'success' | 'error'; error?: string }>(
+    path: string,
+    params: Record<string, string> = {},
+  ): Promise<T> {
    const url = new URL(`${this.prometheusUrl}${path}`);
    Object.entries(params).forEach(([key, value]) => url.searchParams.set(key, value));
-    const response = await fetch(url, { headers: { Accept: 'application/json' }, signal: AbortSignal.timeout(this.queryTimeoutMs) });
+    const response = await fetch(url, {
+      headers: { Accept: 'application/json' },
+      signal: AbortSignal.timeout(this.queryTimeoutMs),
+    });
    if (!response.ok) throw new Error(`Prometheus HTTP ${response.status}`);
-    const result = await response.json() as T;
+    const result = (await response.json()) as T;
    if (result.status !== 'success') throw new Error('Prometheus query failed');
    return result;
  }