Merge branch 'test' into dev

backend/migrations/054_ops_system_logs.sql

+-- 054_ops_system_logs.sql
+-- 统一日志索引表与清理审计表
+
+CREATE TABLE IF NOT EXISTS ops_system_logs (
+  id BIGSERIAL PRIMARY KEY,
+  created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+  level VARCHAR(16) NOT NULL,
+  component VARCHAR(128) NOT NULL DEFAULT '',
+  message TEXT NOT NULL,
+  request_id VARCHAR(128),
+  client_request_id VARCHAR(128),
+  user_id BIGINT,
+  account_id BIGINT,
+  platform VARCHAR(32),
+  model VARCHAR(128),
+  extra JSONB NOT NULL DEFAULT '{}'::jsonb
+);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_created_at_id
+  ON ops_system_logs (created_at DESC, id DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_level_created_at
+  ON ops_system_logs (level, created_at DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_component_created_at
+  ON ops_system_logs (component, created_at DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_request_id
+  ON ops_system_logs (request_id);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_client_request_id
+  ON ops_system_logs (client_request_id);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_user_id_created_at
+  ON ops_system_logs (user_id, created_at DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_account_id_created_at
+  ON ops_system_logs (account_id, created_at DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_platform_model_created_at
+  ON ops_system_logs (platform, model, created_at DESC);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_logs_message_search
+  ON ops_system_logs USING GIN (to_tsvector('simple', COALESCE(message, '')));
+
+CREATE TABLE IF NOT EXISTS ops_system_log_cleanup_audits (
+  id BIGSERIAL PRIMARY KEY,
+  created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+  operator_id BIGINT NOT NULL,
+  conditions JSONB NOT NULL DEFAULT '{}'::jsonb,
+  deleted_rows BIGINT NOT NULL DEFAULT 0
+);
+
+CREATE INDEX IF NOT EXISTS idx_ops_system_log_cleanup_audits_created_at
+  ON ops_system_log_cleanup_audits (created_at DESC, id DESC);

deploy/.env.example

@@ -20,6 +20,52 @@ SERVER_PORT=8080
 # Server mode: release or debug
 SERVER_MODE=release
 
+# -----------------------------------------------------------------------------
+# Logging Configuration
+# 日志配置
+# -----------------------------------------------------------------------------
+# 日志级别：debug/info/warn/error
+LOG_LEVEL=info
+# 日志格式：json/console
+LOG_FORMAT=json
+# 每条日志附带的 service 字段
+LOG_SERVICE_NAME=sub2api
+# 每条日志附带的 env 字段
+LOG_ENV=production
+# 是否输出调用方位置信息
+LOG_CALLER=true
+# 堆栈输出阈值：none/error/fatal
+LOG_STACKTRACE_LEVEL=error
+
+# 输出开关（建议容器内保持双输出）
+# 是否输出到 stdout/stderr
+LOG_OUTPUT_TO_STDOUT=true
+# 是否输出到文件
+LOG_OUTPUT_TO_FILE=true
+# 日志文件路径（留空自动推导）：
+# - 设置 DATA_DIR：${DATA_DIR}/logs/sub2api.log
+# - 未设置 DATA_DIR：/app/data/logs/sub2api.log
+LOG_OUTPUT_FILE_PATH=
+
+# 滚动配置
+# 单文件最大体积（MB）
+LOG_ROTATION_MAX_SIZE_MB=100
+# 保留历史文件数量（0 表示不限制）
+LOG_ROTATION_MAX_BACKUPS=10
+# 历史日志保留天数（0 表示不限制）
+LOG_ROTATION_MAX_AGE_DAYS=7
+# 是否压缩历史日志
+LOG_ROTATION_COMPRESS=true
+# 滚动文件时间戳是否使用本地时间
+LOG_ROTATION_LOCAL_TIME=true
+
+# 采样配置（高频重复日志降噪）
+LOG_SAMPLING_ENABLED=false
+# 每秒前 N 条日志不采样
+LOG_SAMPLING_INITIAL=100
+# 之后每 N 条保留 1 条
+LOG_SAMPLING_THEREAFTER=100
+
 # Global max request body size in bytes (default: 100MB)
 # 全局最大请求体大小（字节，默认 100MB）
 # Applies to all requests, especially important for h2c first request memory protection

deploy/config.example.yaml

@@ -286,6 +286,70 @@ gateway:
     #   profile_2:
     #     name: "Custom Profile 2"
 
+# =============================================================================
+# Logging Configuration
+# 日志配置
+# =============================================================================
+log:
+  # Log level: debug/info/warn/error
+  # 日志级别：debug/info/warn/error
+  level: "info"
+  # Log format: json/console
+  # 日志格式：json/console
+  format: "console"
+  # Service name field written into each log line
+  # 每条日志都会附带 service 字段
+  service_name: "sub2api"
+  # Environment field written into each log line
+  # 每条日志都会附带 env 字段
+  env: "production"
+  # Include caller information
+  # 是否输出调用方位置信息
+  caller: true
+  # Stacktrace threshold: none/error/fatal
+  # 堆栈输出阈值：none/error/fatal
+  stacktrace_level: "error"
+  output:
+    # Keep stdout/stderr output for container log collection
+    # 保持标准输出用于容器日志采集
+    to_stdout: true
+    # Enable file output (default path auto-derived)
+    # 启用文件输出（默认路径自动推导）
+    to_file: true
+    # Empty means:
+    # - DATA_DIR set: {{DATA_DIR}}/logs/sub2api.log
+    # - otherwise: /app/data/logs/sub2api.log
+    # 留空时：
+    # - 设置 DATA_DIR：{{DATA_DIR}}/logs/sub2api.log
+    # - 否则：/app/data/logs/sub2api.log
+    file_path: ""
+  rotation:
+    # Max file size before rotation (MB)
+    # 单文件滚动阈值（MB）
+    max_size_mb: 100
+    # Number of rotated files to keep (0 means unlimited)
+    # 保留历史文件数量（0 表示不限制）
+    max_backups: 10
+    # Number of days to keep old log files (0 means unlimited)
+    # 历史日志保留天数（0 表示不限制）
+    max_age_days: 7
+    # Compress rotated files
+    # 是否压缩历史日志
+    compress: true
+    # Use local time for timestamp in rotated filename
+    # 滚动文件名时间戳使用本地时区
+    local_time: true
+  sampling:
+    # Enable zap sampler (reduce high-frequency repetitive logs)
+    # 启用 zap 采样（减少高频重复日志）
+    enabled: false
+    # Number of first entries per second to always log
+    # 每秒无采样保留的前 N 条日志
+    initial: 100
+    # Thereafter keep 1 out of N entries per second
+    # 之后每 N 条保留 1 条
+    thereafter: 100
+
 # =============================================================================
 # Sora Direct Client Configuration
 # Sora 直连配置

deploy/docker-compose-aicodex.yml

@@ -162,6 +162,10 @@ services:
     volumes:
       - postgres_data:/var/lib/postgresql/data
     environment:
+      # postgres:18-alpine 默认 PGDATA=/var/lib/postgresql/18/docker（位于镜像声明的匿名卷 /var/lib/postgresql 内）。
+      # 若不显式设置 PGDATA，则即使挂载了 postgres_data 到 /var/lib/postgresql/data，数据也不会落盘到该命名卷，
+      # docker compose down/up 后会触发 initdb 重新初始化，导致用户/密码等数据丢失。
+      - PGDATA=/var/lib/postgresql/data
       - POSTGRES_USER=${POSTGRES_USER:-sub2api}
       - POSTGRES_PASSWORD=${POSTGRES_PASSWORD:?POSTGRES_PASSWORD is required}
       - POSTGRES_DB=${POSTGRES_DB:-sub2api}

deploy/docker-compose-test.yml

@@ -142,6 +142,10 @@ services:
     volumes:
       - postgres_data:/var/lib/postgresql/data
     environment:
+      # postgres:18-alpine 默认 PGDATA=/var/lib/postgresql/18/docker（位于镜像声明的匿名卷 /var/lib/postgresql 内）。
+      # 若不显式设置 PGDATA，则即使挂载了 postgres_data 到 /var/lib/postgresql/data，数据也不会落盘到该命名卷，
+      # docker compose down/up 后会触发 initdb 重新初始化，导致用户/密码等数据丢失。
+      - PGDATA=/var/lib/postgresql/data
       - POSTGRES_USER=${POSTGRES_USER:-sub2api}
       - POSTGRES_PASSWORD=${POSTGRES_PASSWORD:?POSTGRES_PASSWORD is required}
       - POSTGRES_DB=${POSTGRES_DB:-sub2api}

deploy/docker-compose.yml

@@ -166,6 +166,10 @@ services:
     volumes:
       - postgres_data:/var/lib/postgresql/data
     environment:
+      # postgres:18-alpine 默认 PGDATA=/var/lib/postgresql/18/docker（位于镜像声明的匿名卷 /var/lib/postgresql 内）。
+      # 若不显式设置 PGDATA，则即使挂载了 postgres_data 到 /var/lib/postgresql/data，数据也不会落盘到该命名卷，
+      # docker compose down/up 后会触发 initdb 重新初始化，导致用户/密码等数据丢失。
+      - PGDATA=/var/lib/postgresql/data
       - POSTGRES_USER=${POSTGRES_USER:-sub2api}
       - POSTGRES_PASSWORD=${POSTGRES_PASSWORD:?POSTGRES_PASSWORD is required}
       - POSTGRES_DB=${POSTGRES_DB:-sub2api}

frontend/src/api/admin/ops.ts

@@ -850,6 +850,77 @@ export interface OpsAggregationSettings {
   aggregation_enabled: boolean
 }
 
+export interface OpsRuntimeLogConfig {
+  level: 'debug' | 'info' | 'warn' | 'error'
+  enable_sampling: boolean
+  sampling_initial: number
+  sampling_thereafter: number
+  caller: boolean
+  stacktrace_level: 'none' | 'error' | 'fatal'
+  retention_days: number
+  source?: string
+  updated_at?: string
+  updated_by_user_id?: number
+}
+
+export interface OpsSystemLog {
+  id: number
+  created_at: string
+  level: string
+  component: string
+  message: string
+  request_id?: string
+  client_request_id?: string
+  user_id?: number | null
+  account_id?: number | null
+  platform?: string
+  model?: string
+  extra?: Record<string, any>
+}
+
+export type OpsSystemLogListResponse = PaginatedResponse<OpsSystemLog>
+
+export interface OpsSystemLogQuery {
+  page?: number
+  page_size?: number
+  time_range?: '5m' | '30m' | '1h' | '6h' | '24h' | '7d' | '30d'
+  start_time?: string
+  end_time?: string
+  level?: string
+  component?: string
+  request_id?: string
+  client_request_id?: string
+  user_id?: number | null
+  account_id?: number | null
+  platform?: string
+  model?: string
+  q?: string
+}
+
+export interface OpsSystemLogCleanupRequest {
+  start_time?: string
+  end_time?: string
+  level?: string
+  component?: string
+  request_id?: string
+  client_request_id?: string
+  user_id?: number | null
+  account_id?: number | null
+  platform?: string
+  model?: string
+  q?: string
+}
+
+export interface OpsSystemLogSinkHealth {
+  queue_depth: number
+  queue_capacity: number
+  dropped_count: number
+  write_failed_count: number
+  written_count: number
+  avg_write_delay_ms: number
+  last_error?: string
+}
+
 export interface OpsErrorLog {
   id: number
   created_at: string
@@ -1205,6 +1276,36 @@ export async function updateAlertRuntimeSettings(config: OpsAlertRuntimeSettings
   return data
 }
 
+export async function getRuntimeLogConfig(): Promise<OpsRuntimeLogConfig> {
+  const { data } = await apiClient.get<OpsRuntimeLogConfig>('/admin/ops/runtime/logging')
+  return data
+}
+
+export async function updateRuntimeLogConfig(config: OpsRuntimeLogConfig): Promise<OpsRuntimeLogConfig> {
+  const { data } = await apiClient.put<OpsRuntimeLogConfig>('/admin/ops/runtime/logging', config)
+  return data
+}
+
+export async function resetRuntimeLogConfig(): Promise<OpsRuntimeLogConfig> {
+  const { data } = await apiClient.post<OpsRuntimeLogConfig>('/admin/ops/runtime/logging/reset')
+  return data
+}
+
+export async function listSystemLogs(params: OpsSystemLogQuery): Promise<OpsSystemLogListResponse> {
+  const { data } = await apiClient.get<OpsSystemLogListResponse>('/admin/ops/system-logs', { params })
+  return data
+}
+
+export async function cleanupSystemLogs(payload: OpsSystemLogCleanupRequest): Promise<{ deleted: number }> {
+  const { data } = await apiClient.post<{ deleted: number }>('/admin/ops/system-logs/cleanup', payload)
+  return data
+}
+
+export async function getSystemLogSinkHealth(): Promise<OpsSystemLogSinkHealth> {
+  const { data } = await apiClient.get<OpsSystemLogSinkHealth>('/admin/ops/system-logs/health')
+  return data
+}
+
 // Advanced settings (DB-backed)
 export async function getAdvancedSettings(): Promise<OpsAdvancedSettings> {
   const { data } = await apiClient.get<OpsAdvancedSettings>('/admin/ops/advanced-settings')
@@ -1272,10 +1373,16 @@ export const opsAPI = {
   updateEmailNotificationConfig,
   getAlertRuntimeSettings,
   updateAlertRuntimeSettings,
+  getRuntimeLogConfig,
+  updateRuntimeLogConfig,
+  resetRuntimeLogConfig,
   getAdvancedSettings,
   updateAdvancedSettings,
   getMetricThresholds,
-  updateMetricThresholds
+  updateMetricThresholds,
+  listSystemLogs,
+  cleanupSystemLogs,
+  getSystemLogSinkHealth
 }
 
 export default opsAPI

frontend/src/views/admin/ops/OpsDashboard.vue

@@ -96,6 +96,13 @@
       <!-- Alert Events -->
       <OpsAlertEventsCard v-if="opsEnabled && !(loading && !hasLoadedOnce)" />
 
+      <!-- System Logs -->
+      <OpsSystemLogTable
+        v-if="opsEnabled && !(loading && !hasLoadedOnce)"
+        :platform-filter="platform"
+        :refresh-token="dashboardRefreshToken"
+      />
+
       <!-- Settings Dialog (hidden in fullscreen mode) -->
       <template v-if="!isFullscreen">
         <OpsSettingsDialog :show="showSettingsDialog" @close="showSettingsDialog = false" @saved="onSettingsSaved" />
@@ -158,6 +165,7 @@ import OpsThroughputTrendChart from './components/OpsThroughputTrendChart.vue'
 import OpsSwitchRateTrendChart from './components/OpsSwitchRateTrendChart.vue'
 import OpsAlertEventsCard from './components/OpsAlertEventsCard.vue'
 import OpsOpenAITokenStatsCard from './components/OpsOpenAITokenStatsCard.vue'
+import OpsSystemLogTable from './components/OpsSystemLogTable.vue'
 import OpsRequestDetailsModal, { type OpsRequestDetailsPreset } from './components/OpsRequestDetailsModal.vue'
 import OpsSettingsDialog from './components/OpsSettingsDialog.vue'
 import OpsAlertRulesCard from './components/OpsAlertRulesCard.vue'

frontend/src/views/admin/ops/components/OpsOpenAITokenStatsCard.vue

@@ -130,8 +130,7 @@ watch(
       next.viewMode !== prev.viewMode ||
       next.pageSize !== prev.pageSize ||
       next.platform !== prev.platform ||
-      next.groupId !== prev.groupId ||
-      next.refreshToken !== prev.refreshToken
+      next.groupId !== prev.groupId
 
     if (next.viewMode === 'pagination' && filtersChanged && next.page !== 1) {
       page.value = 1

frontend/src/views/admin/ops/components/OpsSystemLogTable.vue

+<script setup lang="ts">
+import { computed, onMounted, reactive, ref, watch } from 'vue'
+import { opsAPI, type OpsRuntimeLogConfig, type OpsSystemLog, type OpsSystemLogSinkHealth } from '@/api/admin/ops'
+import Pagination from '@/components/common/Pagination.vue'
+import { useAppStore } from '@/stores'
+
+const appStore = useAppStore()
+
+const props = withDefaults(defineProps<{
+  platformFilter?: string
+  refreshToken?: number
+}>(), {
+  platformFilter: '',
+  refreshToken: 0
+})
+
+const loading = ref(false)
+const logs = ref<OpsSystemLog[]>([])
+const total = ref(0)
+const page = ref(1)
+const pageSize = ref(20)
+
+const health = ref<OpsSystemLogSinkHealth>({
+  queue_depth: 0,
+  queue_capacity: 0,
+  dropped_count: 0,
+  write_failed_count: 0,
+  written_count: 0,
+  avg_write_delay_ms: 0
+})
+
+const runtimeLoading = ref(false)
+const runtimeSaving = ref(false)
+const runtimeConfig = reactive<OpsRuntimeLogConfig>({
+  level: 'info',
+  enable_sampling: false,
+  sampling_initial: 100,
+  sampling_thereafter: 100,
+  caller: true,
+  stacktrace_level: 'error',
+  retention_days: 30
+})
+
+const filters = reactive({
+  time_range: '1h' as '5m' | '30m' | '1h' | '6h' | '24h' | '7d' | '30d',
+  start_time: '',
+  end_time: '',
+  level: '',
+  component: '',
+  request_id: '',
+  client_request_id: '',
+  user_id: '',
+  account_id: '',
+  platform: '',
+  model: '',
+  q: ''
+})
+
+const levelBadgeClass = (level: string) => {
+  const v = String(level || '').toLowerCase()
+  if (v === 'error' || v === 'fatal') return 'bg-red-100 text-red-700 dark:bg-red-900/30 dark:text-red-300'
+  if (v === 'warn' || v === 'warning') return 'bg-amber-100 text-amber-700 dark:bg-amber-900/30 dark:text-amber-300'
+  if (v === 'debug') return 'bg-slate-100 text-slate-700 dark:bg-slate-800 dark:text-slate-300'
+  return 'bg-blue-100 text-blue-700 dark:bg-blue-900/30 dark:text-blue-300'
+}
+
+const formatTime = (value: string) => {
+  if (!value) return '-'
+  const d = new Date(value)
+  if (Number.isNaN(d.getTime())) return value
+  return d.toLocaleString()
+}
+
+const getExtraString = (extra: Record<string, any> | undefined, key: string) => {
+  if (!extra) return ''
+  const v = extra[key]
+  if (v == null) return ''
+  if (typeof v === 'string') return v.trim()
+  if (typeof v === 'number' || typeof v === 'boolean') return String(v)
+  return ''
+}
+
+const formatSystemLogDetail = (row: OpsSystemLog) => {
+  const parts: string[] = []
+  const msg = String(row.message || '').trim()
+  if (msg) parts.push(msg)
+
+  const extra = row.extra || {}
+  const statusCode = getExtraString(extra, 'status_code')
+  const latencyMs = getExtraString(extra, 'latency_ms')
+  const method = getExtraString(extra, 'method')
+  const path = getExtraString(extra, 'path')
+  const clientIP = getExtraString(extra, 'client_ip')
+  const protocol = getExtraString(extra, 'protocol')
+
+  const accessParts: string[] = []
+  if (statusCode) accessParts.push(`status=${statusCode}`)
+  if (latencyMs) accessParts.push(`latency_ms=${latencyMs}`)
+  if (method) accessParts.push(`method=${method}`)
+  if (path) accessParts.push(`path=${path}`)
+  if (clientIP) accessParts.push(`ip=${clientIP}`)
+  if (protocol) accessParts.push(`proto=${protocol}`)
+  if (accessParts.length > 0) parts.push(accessParts.join(' '))
+
+  const corrParts: string[] = []
+  if (row.request_id) corrParts.push(`req=${row.request_id}`)
+  if (row.client_request_id) corrParts.push(`client_req=${row.client_request_id}`)
+  if (row.user_id != null) corrParts.push(`user=${row.user_id}`)
+  if (row.account_id != null) corrParts.push(`acc=${row.account_id}`)
+  if (row.platform) corrParts.push(`platform=${row.platform}`)
+  if (row.model) corrParts.push(`model=${row.model}`)
+  if (corrParts.length > 0) parts.push(corrParts.join(' '))
+
+  const errors = getExtraString(extra, 'errors')
+  if (errors) parts.push(`errors=${errors}`)
+  const err = getExtraString(extra, 'err') || getExtraString(extra, 'error')
+  if (err) parts.push(`error=${err}`)
+
+  // 用空格拼接，交给 CSS 自动换行，尽量“填满再换行”。
+  return parts.join('  ')
+}
+
+const toRFC3339 = (value: string) => {
+  if (!value) return undefined
+  const d = new Date(value)
+  if (Number.isNaN(d.getTime())) return undefined
+  return d.toISOString()
+}
+
+const buildQuery = () => {
+  const query: Record<string, any> = {
+    page: page.value,
+    page_size: pageSize.value,
+    time_range: filters.time_range
+  }
+
+  if (filters.time_range === '30d') {
+    query.time_range = '30d'
+  }
+  if (filters.start_time) query.start_time = toRFC3339(filters.start_time)
+  if (filters.end_time) query.end_time = toRFC3339(filters.end_time)
+  if (filters.level.trim()) query.level = filters.level.trim()
+  if (filters.component.trim()) query.component = filters.component.trim()
+  if (filters.request_id.trim()) query.request_id = filters.request_id.trim()
+  if (filters.client_request_id.trim()) query.client_request_id = filters.client_request_id.trim()
+  if (filters.user_id.trim()) {
+    const v = Number.parseInt(filters.user_id.trim(), 10)
+    if (Number.isFinite(v) && v > 0) query.user_id = v
+  }
+  if (filters.account_id.trim()) {
+    const v = Number.parseInt(filters.account_id.trim(), 10)
+    if (Number.isFinite(v) && v > 0) query.account_id = v
+  }
+  if (filters.platform.trim()) query.platform = filters.platform.trim()
+  if (filters.model.trim()) query.model = filters.model.trim()
+  if (filters.q.trim()) query.q = filters.q.trim()
+  return query
+}
+
+const fetchLogs = async () => {
+  loading.value = true
+  try {
+    const res = await opsAPI.listSystemLogs(buildQuery())
+    logs.value = res.items || []
+    total.value = res.total || 0
+  } catch (err: any) {
+    console.error('[OpsSystemLogTable] Failed to fetch logs', err)
+    appStore.showError(err?.response?.data?.detail || '系统日志加载失败')
+  } finally {
+    loading.value = false
+  }
+}
+
+const fetchHealth = async () => {
+  try {
+    health.value = await opsAPI.getSystemLogSinkHealth()
+  } catch {
+    // 忽略健康数据读取失败，不影响主流程。
+  }
+}
+
+const loadRuntimeConfig = async () => {
+  runtimeLoading.value = true
+  try {
+    const cfg = await opsAPI.getRuntimeLogConfig()
+    runtimeConfig.level = cfg.level
+    runtimeConfig.enable_sampling = cfg.enable_sampling
+    runtimeConfig.sampling_initial = cfg.sampling_initial
+    runtimeConfig.sampling_thereafter = cfg.sampling_thereafter
+    runtimeConfig.caller = cfg.caller
+    runtimeConfig.stacktrace_level = cfg.stacktrace_level
+    runtimeConfig.retention_days = cfg.retention_days
+  } catch (err: any) {
+    console.error('[OpsSystemLogTable] Failed to load runtime log config', err)
+  } finally {
+    runtimeLoading.value = false
+  }
+}
+
+const saveRuntimeConfig = async () => {
+  runtimeSaving.value = true
+  try {
+    const saved = await opsAPI.updateRuntimeLogConfig({ ...runtimeConfig })
+    runtimeConfig.level = saved.level
+    runtimeConfig.enable_sampling = saved.enable_sampling
+    runtimeConfig.sampling_initial = saved.sampling_initial
+    runtimeConfig.sampling_thereafter = saved.sampling_thereafter
+    runtimeConfig.caller = saved.caller
+    runtimeConfig.stacktrace_level = saved.stacktrace_level
+    runtimeConfig.retention_days = saved.retention_days
+    appStore.showSuccess('日志运行时配置已生效')
+  } catch (err: any) {
+    console.error('[OpsSystemLogTable] Failed to save runtime log config', err)
+    appStore.showError(err?.response?.data?.detail || '保存日志配置失败')
+  } finally {
+    runtimeSaving.value = false
+  }
+}
+
+const resetRuntimeConfig = async () => {
+  const ok = window.confirm('确认回滚为启动配置（env/yaml）并立即生效？')
+  if (!ok) return
+
+  runtimeSaving.value = true
+  try {
+    const saved = await opsAPI.resetRuntimeLogConfig()
+    runtimeConfig.level = saved.level
+    runtimeConfig.enable_sampling = saved.enable_sampling
+    runtimeConfig.sampling_initial = saved.sampling_initial
+    runtimeConfig.sampling_thereafter = saved.sampling_thereafter
+    runtimeConfig.caller = saved.caller
+    runtimeConfig.stacktrace_level = saved.stacktrace_level
+    runtimeConfig.retention_days = saved.retention_days
+    appStore.showSuccess('已回滚到启动日志配置')
+    await fetchHealth()
+  } catch (err: any) {
+    console.error('[OpsSystemLogTable] Failed to reset runtime log config', err)
+    appStore.showError(err?.response?.data?.detail || '回滚日志配置失败')
+  } finally {
+    runtimeSaving.value = false
+  }
+}
+
+const cleanupCurrentFilter = async () => {
+  const ok = window.confirm('确认按当前筛选条件清理系统日志？该操作不可撤销。')
+  if (!ok) return
+  try {
+    const payload = {
+      start_time: toRFC3339(filters.start_time),
+      end_time: toRFC3339(filters.end_time),
+      level: filters.level.trim() || undefined,
+      component: filters.component.trim() || undefined,
+      request_id: filters.request_id.trim() || undefined,
+      client_request_id: filters.client_request_id.trim() || undefined,
+      user_id: filters.user_id.trim() ? Number.parseInt(filters.user_id.trim(), 10) : undefined,
+      account_id: filters.account_id.trim() ? Number.parseInt(filters.account_id.trim(), 10) : undefined,
+      platform: filters.platform.trim() || undefined,
+      model: filters.model.trim() || undefined,
+      q: filters.q.trim() || undefined
+    }
+    const res = await opsAPI.cleanupSystemLogs(payload)
+    appStore.showSuccess(`清理完成，删除 ${res.deleted || 0} 条日志`)
+    page.value = 1
+    await Promise.all([fetchLogs(), fetchHealth()])
+  } catch (err: any) {
+    console.error('[OpsSystemLogTable] Failed to cleanup logs', err)
+    appStore.showError(err?.response?.data?.detail || '清理系统日志失败')
+  }
+}
+
+const resetFilters = () => {
+  filters.time_range = '1h'
+  filters.start_time = ''
+  filters.end_time = ''
+  filters.level = ''
+  filters.component = ''
+  filters.request_id = ''
+  filters.client_request_id = ''
+  filters.user_id = ''
+  filters.account_id = ''
+  filters.platform = props.platformFilter || ''
+  filters.model = ''
+  filters.q = ''
+  page.value = 1
+  fetchLogs()
+}
+
+watch(() => props.platformFilter, (v) => {
+  if (v && !filters.platform) {
+    filters.platform = v
+    page.value = 1
+    fetchLogs()
+  }
+})
+
+watch(() => props.refreshToken, () => {
+  fetchLogs()
+  fetchHealth()
+})
+
+const onPageChange = (next: number) => {
+  page.value = next
+  fetchLogs()
+}
+
+const onPageSizeChange = (next: number) => {
+  pageSize.value = next
+  page.value = 1
+  fetchLogs()
+}
+
+const applyFilters = () => {
+  page.value = 1
+  fetchLogs()
+}
+
+const hasData = computed(() => logs.value.length > 0)
+
+onMounted(async () => {
+  if (props.platformFilter) {
+    filters.platform = props.platformFilter
+  }
+  await Promise.all([fetchLogs(), fetchHealth(), loadRuntimeConfig()])
+})
+</script>
+
+<template>
+  <section class="rounded-2xl border border-gray-200 bg-white p-4 shadow-sm dark:border-dark-700 dark:bg-dark-900/60">
+    <div class="mb-4 flex flex-wrap items-center justify-between gap-3">
+      <div>
+        <h3 class="text-sm font-bold text-gray-900 dark:text-white">系统日志</h3>
+        <p class="mt-1 text-xs text-gray-500 dark:text-gray-400">默认按最新时间倒序，支持筛选搜索与按条件清理。</p>
+      </div>
+      <div class="flex flex-wrap items-center gap-2 text-xs">
+        <span class="rounded-md bg-gray-100 px-2 py-1 text-gray-700 dark:bg-dark-700 dark:text-gray-200">队列 {{ health.queue_depth }}/{{ health.queue_capacity }}</span>
+        <span class="rounded-md bg-gray-100 px-2 py-1 text-gray-700 dark:bg-dark-700 dark:text-gray-200">写入 {{ health.written_count }}</span>
+        <span class="rounded-md bg-amber-100 px-2 py-1 text-amber-700 dark:bg-amber-900/30 dark:text-amber-300">丢弃 {{ health.dropped_count }}</span>
+        <span class="rounded-md bg-red-100 px-2 py-1 text-red-700 dark:bg-red-900/30 dark:text-red-300">失败 {{ health.write_failed_count }}</span>
+      </div>
+    </div>
+
+    <div class="mb-4 rounded-xl border border-gray-200 bg-gray-50 p-3 dark:border-dark-700 dark:bg-dark-800/70">
+      <div class="mb-2 flex items-center justify-between">
+        <div class="text-xs font-semibold text-gray-700 dark:text-gray-200">运行时日志配置（实时生效）</div>
+        <span v-if="runtimeLoading" class="text-xs text-gray-500">加载中...</span>
+      </div>
+      <div class="grid grid-cols-1 gap-3 md:grid-cols-6">
+        <label class="text-xs text-gray-600 dark:text-gray-300">
+          级别
+          <select v-model="runtimeConfig.level" class="input mt-1">
+            <option value="debug">debug</option>
+            <option value="info">info</option>
+            <option value="warn">warn</option>
+            <option value="error">error</option>
+          </select>
+        </label>
+        <label class="text-xs text-gray-600 dark:text-gray-300">
+          堆栈阈值
+          <select v-model="runtimeConfig.stacktrace_level" class="input mt-1">
+            <option value="none">none</option>
+            <option value="error">error</option>
+            <option value="fatal">fatal</option>
+          </select>
+        </label>
+        <label class="text-xs text-gray-600 dark:text-gray-300">
+          采样初始
+          <input v-model.number="runtimeConfig.sampling_initial" type="number" min="1" class="input mt-1" />
+        </label>
+        <label class="text-xs text-gray-600 dark:text-gray-300">
+          采样后续
+          <input v-model.number="runtimeConfig.sampling_thereafter" type="number" min="1" class="input mt-1" />
+        </label>
+        <label class="text-xs text-gray-600 dark:text-gray-300">
+          保留天数
+          <input v-model.number="runtimeConfig.retention_days" type="number" min="1" max="3650" class="input mt-1" />
+        </label>
+        <div class="flex items-end gap-2">
+          <label class="inline-flex items-center gap-2 text-xs text-gray-600 dark:text-gray-300">
+            <input v-model="runtimeConfig.caller" type="checkbox" />
+            caller
+          </label>
+          <label class="inline-flex items-center gap-2 text-xs text-gray-600 dark:text-gray-300">
+            <input v-model="runtimeConfig.enable_sampling" type="checkbox" />
+            sampling
+          </label>
+          <button type="button" class="btn btn-primary btn-sm" :disabled="runtimeSaving" @click="saveRuntimeConfig">
+            {{ runtimeSaving ? '保存中...' : '保存并生效' }}
+          </button>
+          <button type="button" class="btn btn-secondary btn-sm" :disabled="runtimeSaving" @click="resetRuntimeConfig">
+            回滚默认值
+          </button>
+        </div>
+      </div>
+      <p v-if="health.last_error" class="mt-2 text-xs text-red-600 dark:text-red-400">最近写入错误：{{ health.last_error }}</p>
+    </div>
+
+    <div class="mb-4 grid grid-cols-1 gap-3 md:grid-cols-5">
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        时间范围
+        <select v-model="filters.time_range" class="input mt-1">
+          <option value="5m">5m</option>
+          <option value="30m">30m</option>
+          <option value="1h">1h</option>
+          <option value="6h">6h</option>
+          <option value="24h">24h</option>
+          <option value="7d">7d</option>
+          <option value="30d">30d</option>
+        </select>
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        开始时间（可选）
+        <input v-model="filters.start_time" type="datetime-local" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        结束时间（可选）
+        <input v-model="filters.end_time" type="datetime-local" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        级别
+        <select v-model="filters.level" class="input mt-1">
+          <option value="">全部</option>
+          <option value="debug">debug</option>
+          <option value="info">info</option>
+          <option value="warn">warn</option>
+          <option value="error">error</option>
+        </select>
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        组件
+        <input v-model="filters.component" type="text" class="input mt-1" placeholder="如 http.access" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        request_id
+        <input v-model="filters.request_id" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        client_request_id
+        <input v-model="filters.client_request_id" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        user_id
+        <input v-model="filters.user_id" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        account_id
+        <input v-model="filters.account_id" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        平台
+        <input v-model="filters.platform" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        模型
+        <input v-model="filters.model" type="text" class="input mt-1" />
+      </label>
+      <label class="text-xs text-gray-600 dark:text-gray-300">
+        关键词
+        <input v-model="filters.q" type="text" class="input mt-1" placeholder="消息/request_id" />
+      </label>
+    </div>
+
+    <div class="mb-3 flex flex-wrap gap-2">
+      <button type="button" class="btn btn-primary btn-sm" @click="applyFilters">查询</button>
+      <button type="button" class="btn btn-secondary btn-sm" @click="resetFilters">重置</button>
+      <button type="button" class="btn btn-danger btn-sm" @click="cleanupCurrentFilter">按当前筛选清理</button>
+      <button type="button" class="btn btn-secondary btn-sm" @click="fetchHealth">刷新健康指标</button>
+    </div>
+
+    <div class="overflow-hidden rounded-xl border border-gray-200 dark:border-dark-700">
+      <div v-if="loading" class="px-4 py-8 text-center text-sm text-gray-500">加载中...</div>
+      <div v-else-if="!hasData" class="px-4 py-8 text-center text-sm text-gray-500">暂无系统日志</div>
+      <div v-else class="overflow-auto">
+        <table class="min-w-full table-fixed divide-y divide-gray-200 dark:divide-dark-700">
+          <thead class="bg-gray-50 dark:bg-dark-900">
+            <tr>
+              <th class="w-[170px] px-3 py-2 text-left text-[11px] font-semibold text-gray-500">时间</th>
+              <th class="w-[80px] px-3 py-2 text-left text-[11px] font-semibold text-gray-500">级别</th>
+              <th class="px-3 py-2 text-left text-[11px] font-semibold text-gray-500">日志详细信息</th>
+            </tr>
+          </thead>
+          <tbody class="divide-y divide-gray-100 dark:divide-dark-800">
+            <tr v-for="row in logs" :key="row.id" class="align-top">
+              <td class="px-3 py-2 text-xs text-gray-700 dark:text-gray-300">{{ formatTime(row.created_at) }}</td>
+              <td class="px-3 py-2 text-xs">
+                <span class="inline-flex rounded-full px-2 py-0.5 font-semibold" :class="levelBadgeClass(row.level)">
+                  {{ row.level }}
+                </span>
+              </td>
+              <td class="px-3 py-2 text-xs text-gray-700 dark:text-gray-300 whitespace-normal break-all">
+                {{ formatSystemLogDetail(row) }}
+              </td>
+            </tr>
+          </tbody>
+        </table>
+      </div>
+      <Pagination
+        :total="total"
+        :page="page"
+        :page-size="pageSize"
+        :page-size-options="[10, 20, 50, 100, 200]"
+        @update:page="onPageChange"
+        @update:page-size="onPageSizeChange"
+      />
+    </div>
+  </section>
+</template>

frontend/src/views/admin/ops/types.ts

@@ -17,5 +17,8 @@ export type {
   OpsMetricThresholds,
   OpsAdvancedSettings,
   OpsDataRetentionSettings,
-  OpsAggregationSettings
+  OpsAggregationSettings,
+  OpsRuntimeLogConfig,
+  OpsSystemLog,
+  OpsSystemLogSinkHealth
 } from '@/api/admin/ops'

logging_audit_20260212.md

+# 日志专项审计与整理（2026-02-12）
+
+## 1. 全量扫描结论
+
+- 扫描范围：`backend/` + `frontend/`
+- 日志相关调用总量（粗统计）：约 `4100` 处
+- 后端标准库日志（`log.Printf/Println/Fatal*`）：`808` 处（本轮整改后剩余 `269` 处）
+- 前端 `console.*`：`180` 处
+
+关键观察：
+
+1. 后端大量业务日志仍走标准库 `log`，在当前初始化流程里会被统一当作 `INFO` 输出，导致“错误/告警等级失真”。
+2. 网关关键链路（OpenAI/Gemini/Sora）原有日志以格式化字符串为主，上下文字段（`request_id/user_id/group_id/model/account_id`）不完整，排障时需要人工拼接上下文。
+3. Token 刷新服务同时混用 `log` 与 `slog`，同类事件日志风格不一致，不利于检索与聚合。
+4. 前端 `console.error/warn` 使用量高，缺少统一封装，生产环境噪音和敏感信息泄漏风险较高。
+
+## 2. 本次已落地整改
+
+### 2.1 全局层（后端标准库日志分级修复）
+
+- 修改：`backend/internal/pkg/logger/logger.go`
+- 结果：
+  1. 替换原 `zap.RedirectStdLogAt(..., INFO)` 机制，改为自定义 `stdlog bridge`。
+  2. 对标准库日志自动推断等级（`DEBUG/WARN/ERROR/INFO`），并打上 `legacy_stdlog=true` 标记。
+  3. 规范化消息文本（去换行、压缩空白），提升可读性和检索稳定性。
+  4. 调整初始化顺序：先桥接 `slog`，再桥接 `stdlog`，避免 `slog.SetDefault` 覆盖标准库桥接。
+  5. 新增 `logger.LegacyPrintf(component, format, ...args)`，用于后端历史 `printf` 日志的平滑迁移，自动推断等级并打 `legacy_printf=true` 标记。
+
+### 2.2 核心请求链路结构化改造
+
+- 新增：`backend/internal/handler/logging.go`
+  - 统一提供请求级 logger 获取入口，继承中间件注入的 `request_id` 上下文。
+
+- 改造文件：
+  - `backend/internal/handler/gateway_handler.go`
+  - `backend/internal/handler/openai_gateway_handler.go`
+  - `backend/internal/handler/gemini_v1beta_handler.go`
+  - `backend/internal/handler/sora_gateway_handler.go`
+  - `backend/internal/service/antigravity_gateway_service.go`
+  - `backend/internal/service/gateway_service.go`
+  - `backend/internal/service/gemini_oauth_service.go`
+  - `backend/internal/service/auth_service.go`
+  - `backend/internal/setup/setup.go`
+  - `backend/internal/service/usage_cleanup_service.go`
+  - `backend/internal/service/pricing_service.go`
+  - `backend/internal/repository/account_repo.go`
+  - `backend/internal/service/openai_gateway_service.go`
+  - `backend/internal/service/scheduler_snapshot_service.go`
+  - `backend/internal/service/gemini_messages_compat_service.go`
+  - `backend/internal/service/dashboard_aggregation_service.go`
+  - `backend/internal/service/billing_cache_service.go`
+  - `backend/internal/repository/claude_oauth_service.go`
+  - `backend/internal/service/admin_service.go`
+  - `backend/internal/handler/admin/ops_ws_handler.go`
+
+- 改造内容：
+  1. 把关键日志从字符串拼接改为结构化字段。
+  2. 统一带上 `component/user_id/api_key_id/group_id/model/account_id` 等字段。
+  3. 按语义拆分等级：
+     - 预期业务拒绝（如账单校验失败、队列满）使用 `Info`
+     - 降级路径/可恢复异常（如抢槽失败、粘性会话绑定失败）使用 `Warn`
+     - 真正故障（如转发失败、使用量记录失败）使用 `Error`
+  4. 新增请求完成日志（`*.request_completed`）用于链路闭环追踪。
+  5. 对高密度 `log.Printf` 完成批量迁移到 `logger.LegacyPrintf`（本轮累计 511 处），并统一组件字段：
+     - `component=service.antigravity_gateway`
+     - `component=service.gateway`
+     - `component=service.gemini_oauth`
+     - `component=service.auth`
+     - `component=setup`
+     - `component=service.usage_cleanup`
+     - `component=service.pricing`
+     - `component=repository.account`
+     - `component=service.openai_gateway`
+     - `component=service.scheduler_snapshot`
+     - `component=service.gemini_messages_compat`
+     - `component=service.dashboard_aggregation`
+     - `component=service.billing_cache`
+     - `component=repository.claude_oauth`
+     - `component=service.admin`
+     - `component=handler.admin.ops_ws`
+  6. OpenAI 透传断流相关两条关键告警统一回到新日志系统输出（`service.openai_gateway`），并通过兼容逻辑保证测试环境可捕获。
+
+### 2.3 后台任务日志统一
+
+- 改造：`backend/internal/service/token_refresh_service.go`
+- 结果：
+  1. 统一改为 `slog` 结构化输出。
+  2. `retry/cycle/account` 等事件改为字段化日志，便于按账号和批次检索。
+  3. 对“无实际刷新活动”的周期日志降级到 `Debug`，减少噪音。
+
+### 2.4 测试保障
+
+- 新增：`backend/internal/pkg/logger/stdlog_bridge_test.go`
+  - 覆盖标准库日志等级推断、消息标准化、输出路由行为。
+- 已验证：
+  - `go test ./internal/pkg/logger ./internal/handler ./internal/service` 通过。
+
+## 3. 仍需继续整改（建议下一批）
+
+### 3.1 后端剩余 `std log` 高密度区域（优先级 P1）
+
+建议优先处理以下文件（调用量高）：
+
+1. `backend/internal/service/usage_cleanup_service.go`（26）
+2. `backend/internal/service/pricing_service.go`（26）
+3. `backend/internal/repository/account_repo.go`（24）
+4. `backend/internal/service/openai_gateway_service.go`（23）
+5. `backend/internal/service/scheduler_snapshot_service.go`（20）
+
+（以上已完成。当前 Top 5 已变为：`backend/cmd/server/main.go`、`backend/internal/service/openai_tool_corrector.go`、`backend/internal/service/email_queue_service.go`、`backend/internal/config/config.go`、`backend/internal/service/ops_cleanup_service.go`）
+
+目标：逐步替换为结构化日志，减少对 `legacy_stdlog` 兼容桥接的依赖。
+
+### 3.2 前端日志治理（优先级 P1）
+
+建议新增统一前端日志工具（如 `src/utils/logger.ts`）并分三步替换：
+
+1. `console.error/warn/debug/log` 全部收敛到统一 API；
+2. 生产环境默认降噪（仅保留关键告警/错误）；
+3. 统一字段（模块名、请求ID、用户ID、路由、错误码）并避免打印敏感数据。
+
+### 3.3 日志规范与门禁（优先级 P2）
+
+建议补充：
+
+1. 日志规范文档（等级定义、字段最小集、脱敏要求）；
+2. CI 检查规则：限制新增裸 `log.Printf` / `console.*`；
+3. 面向运营告警的事件白名单（例如 `*.forward_failed`、`*.retry_exhausted*`）。
+
+## 4. 本次整理后可直接使用的检索建议
+
+1. 过滤历史兼容日志：`legacy_stdlog=true`
+2. 网关入口故障：`component=handler.* AND level in (WARN,ERROR)`
+3. 请求闭环：按 `request_id` + `*.request_completed` + `*.forward_failed`
+4. token 刷新故障：`component=*token_refresh* AND (retry_attempt_failed OR set_error_status_failed)`