hailin
893513ad78
feat(presence): 添加设备档案表,实现事件流水+设备快照分离
...
大厂标准架构(Amplitude/Mixpanel):
- analytics_event_log: 事件流水(append-only,每条事件一行)
- analytics_device_profile: 设备快照(每台设备一行,upsert 更新)
设备分布查询从 O(events) 降为 O(devices):
- SELECT COUNT(*), device_brand FROM analytics_device_profile GROUP BY device_brand
不再需要 COUNT(DISTINCT install_id) 扫描全量事件表
ON CONFLICT (install_id) DO UPDATE:
- COALESCE 保留已有字段(不被 NULL 覆盖)
- last_seen_at 每次上报更新
- event_count 累加(可用于活跃度分析)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-05 19:31:33 -08:00
hailin
6bca65e434
feat(telemetry): 设备字段提升为顶层结构化列(Amplitude 风格)
...
将 device_brand/device_model/device_os/app_version/locale 从 JSONB properties
提升为 analytics_event_log 表的独立列,并建立索引,支持亿级数据量下的高效
按设备维度查询和分组统计。
前端 (mining-app + mobile-app):
- toServerJson() 从 properties 中提取设备字段,以顶层字段发送给服务端
- 本地存储格式不变(properties 仍保留设备字段,便于离线队列完整性)
后端 (presence-service):
- Prisma schema: EventLog 新增 deviceBrand/deviceModel/deviceOs/appVersion/locale 列
- Migration: ALTER TABLE 添加 5 列 + 2 个索引
- DTO/Command: EventItemDto 接收顶层设备字段
- Entity: EventLog 新增 5 个字段及 getter
- Mapper: toDomain/toPersistence 映射新字段
- Handler: toEventLog 从 DTO 读取设备字段;SessionStartedEvent 优先使用顶层字段
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-05 09:50:48 -08:00
hailin
e8d9cb72a9
feat(presence-service): 修复鉴权——用户JWT验证心跳,管理员JWT查询在线/DAU数据
...
- 添加 @nestjs/jwt 依赖,AppModule 注册全局 JwtModule
- 重写 JwtAuthGuard:使用 JwtService.verifyAsync 解析用户 token (type=access)
- 新建 AdminGuard:验证管理员 token (type=admin),与 identity-service 共享 JWT_SECRET
- heartbeat 接口:保持 JwtAuthGuard(用户 JWT)
- online-count / online-history / dau:改用 AdminGuard(管理员 JWT)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-05 08:52:39 -08:00
hailin
cc06638e0e
feat(telemetry): 将userId改为userSerialNum字符串格式并完善遥测追踪
...
Backend (presence-service):
- 将EventLog.userId从BigInt改为String类型,存储userSerialNum(如D25121400005)
- 更新Prisma schema,userId字段改为VarChar(20)并添加索引
- 更新心跳相关命令和事件,统一使用userSerialNum字符串
- 添加数据库迁移文件
- 更新相关单元测试和集成测试
Frontend (mobile-app):
- TelemetryEvent新增toServerJson()方法,格式化为后端API期望的格式
- AccountService登录/恢复时设置TelemetryService的userId
- MultiAccountService切换账号时同步更新TelemetryService的userId
- 退出登录时清除TelemetryService的userId
- AuthProvider初始化时设置userId
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-15 06:55:25 -08:00
hailin
747e4ae8ef
refactor(mpc-system): migrate to party-driven architecture with PartyID-based routing
...
- Remove Address field from PartyEndpoint (parties connect to router themselves)
- Update K8s Discovery to only manage PartyID and Role labels
- Add Party registration and SessionEvent protobuf definitions
- Implement PartyRegistry and SessionEventBroadcaster domain logic
- Add RegisterParty and SubscribeSessionEvents gRPC handlers
- Prepare infrastructure for party-driven MPC coordination
This is the first phase of migrating from coordinator-driven to party-driven
architecture following international MPC system design patterns.
2025-12-05 08:11:28 -08:00
Developer
f771dc8f6e
fix(presence-service): 修复 MetricsModule 依赖注入,直接提供仓库实现
...
MetricsCollectorService 需要 DAU 和 EventLog 仓库,
但这些在 InfrastructureModule 中通过符号注入,
导致循环依赖问题。改为在 MetricsModule 中直接提供这些依赖。
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-04 01:35:59 -08:00
Developer
8faaedaa64
fix(presence-service): MetricsModule 导入 RedisModule 和 PersistenceModule
...
修复依赖注入错误: PresenceRedisRepository 和 DAU/EventLog 仓库无法解析
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-04 01:30:08 -08:00
Developer
43e769437f
fix(presence-service): 修复方法名 findByDay -> findByDate
...
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-04 00:53:04 -08:00
Developer
9656a6f4c4
feat(presence-service): 添加 Prometheus 指标监控和 Grafana Dashboard
...
Prometheus 指标端点 (/api/v1/metrics):
- presence_online_users_total: 实时在线人数
- presence_dau_total: 今日 DAU
- presence_heartbeat_total: 心跳计数
- presence_events_received_total: 事件上报计数
- presence_session_start_total: 会话开始计数
- presence_heartbeat_duration_seconds: 心跳处理延迟
- presence_event_batch_duration_seconds: 事件批处理延迟
Grafana Dashboard:
- 核心指标概览 (在线人数、DAU、心跳、事件)
- 趋势图表 (在线人数趋势、心跳/事件速率)
- 事件分布 (饼图、按小时趋势)
- 性能指标 (P50/P95/P99 延迟)
- 服务资源 (内存、CPU)
配置更新:
- prometheus.yml 添加 presence-service 抓取配置
- package.json 添加 prom-client 依赖
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-03 08:38:05 -08:00
Developer
086d82f98a
feat(admin-service): 添加 init-db 命令自动创建数据库
...
用法: ./deploy.sh init-db
- 自动创建 rwa_admin 数据库(如不存在)
- 自动运行 Prisma 迁移
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-02 22:02:41 -08:00
Developer
603c1c6c90
feat(presence-service): 添加完整的测试套件和生产部署设施
...
- 添加 165 个测试用例 (123 单元测试, 22 集成测试, 20 E2E测试)
- 添加 Docker 多阶段构建和 docker-compose 生产部署配置
- 添加完整的文档体系 (架构、API、开发、测试、部署)
- 添加环境配置 (.env.development/production/test)
- 添加部署脚本 (health-check, start-all, stop-service)
- 修复 TypeScript 类型错误
- 经 WSL2 验证所有生产部署命令测试通过
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-02 20:10:04 -08:00
Developer
0be3fe619e
feat(presence-service): 完整实现 DDD+Hexagonal 架构的用户活跃度与在线状态服务
...
Domain 层 (领域层):
- 值对象: InstallId, EventName, EventProperties, DeviceInfo, TimeWindow
- 实体: EventLog, OnlineSnapshot
- 聚合根: DailyActiveStats
- 领域事件: SessionStartedEvent, HeartbeatReceivedEvent, DauCalculatedEvent
- 仓储接口: IEventLogRepository, IDailyActiveStatsRepository, IOnlineSnapshotRepository
- 领域服务: DauCalculationService, OnlineDetectionService
Infrastructure 层 (基础设施层):
- Prisma: EventLog, DailyActiveStats, OnlineSnapshot 数据模型
- Redis: 在线状态存储 (ZSET) + HyperLogLog DAU 实时统计
- Kafka: 事件发布服务 (可选)
- Mappers: 领域对象 <-> Prisma 模型转换
- 仓储实现: EventLogRepositoryImpl, DailyActiveStatsRepositoryImpl, OnlineSnapshotRepositoryImpl
Application 层 (应用层):
- Commands: RecordEvents, RecordHeartbeat, CalculateDau
- Queries: GetOnlineCount, GetDauStats
- Schedulers: 每分钟记录在线快照, 每小时清理过期数据, 每天凌晨计算前一天DAU
API 层 (表现层):
- Controllers: AnalyticsController, PresenceController, HealthController
- DTOs: BatchEventsDto, HeartbeatDto, QueryDauDto, OnlineCountResponseDto, DauStatsResponseDto
- Guards: JwtAuthGuard
- Decorators: @Public, @CurrentUser
核心功能:
- ✅ 用户行为事件批量上报与存储
- ✅ 日活 DAU 统计 (按自然日去重, 支持省/市维度)
- ✅ 实时在线人数统计 (3分钟窗口)
- ✅ 心跳机制 (前台60秒间隔)
- ✅ HyperLogLog 实时 DAU 计数
- ✅ 定时任务 (快照记录, 过期清理, DAU 计算)
- ✅ Swagger API 文档
- ✅ Docker 多阶段构建
- ✅ 健康检查
技术栈:
- NestJS + TypeScript
- Prisma ORM + PostgreSQL
- Redis (ioredis)
- Kafka (kafkajs, 可选)
- CQRS 模式 (@nestjs/cqrs)
- 定时任务 (@nestjs/schedule)
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-12-02 12:11:38 -08:00