Version: 0.9.68.dev.260504

后端：
1. 阶段 3 notification 服务边界落地，新增 `cmd/notification`、`services/notification`、`gateway/notification`、`shared/contracts/notification` 和 notification port，按 userauth 同款最小手搓 zrpc 样板收口
2. notification outbox consumer、relay 和 retry loop 迁入独立服务入口，处理 `notification.feishu.requested`，gateway 改为通过 zrpc client 调用 notification
3. 清退旧单体 notification DAO/model/service/provider/runner 和 `service/events/notification_feishu.go`，旧实现不再作为活跃编译路径
4. 修复 outbox 路由归属、dispatch 启动扫描、Kafka topic 探测/投递超时、sending 租约恢复、毒消息 MarkDead 错误回传和 RPC timeout 边界
5. 同步调整 active-scheduler 触发通知事件、核心 outbox handler、MySQL 迁移边界和 notification 配置

文档：
1. 更新微服务迁移计划，将阶段 3 notification 标记为已完成，并明确下一阶段从 active-scheduler 开始

backend/services/notification/dao/record.go

@@ -0,0 +1,170 @@
+package dao
+
+import (
+	"context"
+	"errors"
+	"time"
+
+	notificationmodel "github.com/LoveLosita/smartflow/backend/services/notification/model"
+	"gorm.io/gorm"
+)
+
+// RecordDAO 管理 notification_records 投递状态机持久化。
+//
+// 职责边界：
+// 1. 只负责通知记录的创建、去重查询、状态更新和重试扫描；
+// 2. 不负责 provider 发送、幂等锁或 outbox consumed 标记；
+// 3. 不读写 active_schedule_* 表，避免 notification 服务反向持有主动调度内部状态。
+type RecordDAO struct {
+	db *gorm.DB
+}
+
+func NewRecordDAO(db *gorm.DB) *RecordDAO {
+	return &RecordDAO{db: db}
+}
+
+func (d *RecordDAO) WithTx(tx *gorm.DB) *RecordDAO {
+	return &RecordDAO{db: tx}
+}
+
+func (d *RecordDAO) ensureDB() error {
+	if d == nil || d.db == nil {
+		return errors.New("notification record dao 未初始化")
+	}
+	return nil
+}
+
+func (d *RecordDAO) CreateNotificationRecord(ctx context.Context, record *notificationmodel.NotificationRecord) error {
+	if err := d.ensureDB(); err != nil {
+		return err
+	}
+	if record == nil {
+		return errors.New("notification record 不能为空")
+	}
+	return d.db.WithContext(ctx).Create(record).Error
+}
+
+func (d *RecordDAO) UpdateNotificationRecordFields(ctx context.Context, notificationID int64, updates map[string]any) error {
+	if err := d.ensureDB(); err != nil {
+		return err
+	}
+	if notificationID <= 0 {
+		return errors.New("notification record id 不能为空")
+	}
+	if len(updates) == 0 {
+		return nil
+	}
+	return d.db.WithContext(ctx).
+		Model(&notificationmodel.NotificationRecord{}).
+		Where("id = ?", notificationID).
+		Updates(updates).Error
+}
+
+func (d *RecordDAO) GetNotificationRecordByID(ctx context.Context, notificationID int64) (*notificationmodel.NotificationRecord, error) {
+	if err := d.ensureDB(); err != nil {
+		return nil, err
+	}
+	if notificationID <= 0 {
+		return nil, gorm.ErrRecordNotFound
+	}
+	var record notificationmodel.NotificationRecord
+	err := d.db.WithContext(ctx).Where("id = ?", notificationID).First(&record).Error
+	if err != nil {
+		return nil, err
+	}
+	return &record, nil
+}
+
+// FindNotificationRecordByDedupeKey 查询通知去重记录。
+//
+// 说明：
+// 1. notification 第一版按 channel + dedupe_key 聚合去重；
+// 2. 若返回 pending/sending/sent，上层应避免重复投递；
+// 3. 若返回 failed，上层可以复用同一条记录进入 provider retry。
+func (d *RecordDAO) FindNotificationRecordByDedupeKey(ctx context.Context, channel string, dedupeKey string) (*notificationmodel.NotificationRecord, error) {
+	if err := d.ensureDB(); err != nil {
+		return nil, err
+	}
+	if channel == "" || dedupeKey == "" {
+		return nil, gorm.ErrRecordNotFound
+	}
+	var record notificationmodel.NotificationRecord
+	err := d.db.WithContext(ctx).
+		Where("channel = ? AND dedupe_key = ?", channel, dedupeKey).
+		Order("created_at DESC, id DESC").
+		First(&record).Error
+	if err != nil {
+		return nil, err
+	}
+	return &record, nil
+}
+
+// ListRetryableNotificationRecords 查询到达重试时间的通知记录。
+//
+// 1. failed 记录按 next_retry_at 进入重试队列；
+// 2. sending 记录只有超过租约才会回收，避免仍在执行的 provider 调用被重复放大；
+// 3. 这让 retry scanner 同时覆盖显式失败重试和“发送中崩溃恢复”。
+func (d *RecordDAO) ListRetryableNotificationRecords(ctx context.Context, now time.Time, sendingStaleBefore time.Time, limit int) ([]notificationmodel.NotificationRecord, error) {
+	if err := d.ensureDB(); err != nil {
+		return nil, err
+	}
+	if limit <= 0 || now.IsZero() {
+		return []notificationmodel.NotificationRecord{}, nil
+	}
+	if sendingStaleBefore.IsZero() {
+		sendingStaleBefore = now.Add(-10 * time.Minute)
+	}
+	var records []notificationmodel.NotificationRecord
+	err := d.db.WithContext(ctx).
+		Where(
+			"(status = ? AND next_retry_at IS NOT NULL AND next_retry_at <= ?) OR (status = ? AND updated_at <= ?)",
+			notificationmodel.RecordStatusFailed,
+			now,
+			notificationmodel.RecordStatusSending,
+			sendingStaleBefore,
+		).
+		Order("next_retry_at ASC, id ASC").
+		Limit(limit).
+		Find(&records).Error
+	if err != nil {
+		return nil, err
+	}
+	return records, nil
+}
+
+// ClaimRetryableNotificationRecord 抢占一条到期失败通知，避免多实例重复调用 provider。
+//
+// 职责边界：
+// 1. 只做跨进程 claim，不发送通知、不推进最终投递状态；
+// 2. failed 到期记录和 stale sending 记录都可以被回收为 sending；
+// 3. 返回 claimed=false 表示记录已被其它实例抢走或状态已变化，调用方应跳过本次重试。
+func (d *RecordDAO) ClaimRetryableNotificationRecord(ctx context.Context, notificationID int64, now time.Time, sendingStaleBefore time.Time) (bool, error) {
+	if err := d.ensureDB(); err != nil {
+		return false, err
+	}
+	if notificationID <= 0 || now.IsZero() {
+		return false, nil
+	}
+	if sendingStaleBefore.IsZero() {
+		sendingStaleBefore = now.Add(-10 * time.Minute)
+	}
+	result := d.db.WithContext(ctx).
+		Model(&notificationmodel.NotificationRecord{}).
+		Where(
+			"id = ? AND ((status = ? AND next_retry_at IS NOT NULL AND next_retry_at <= ?) OR (status = ? AND updated_at <= ?))",
+			notificationID,
+			notificationmodel.RecordStatusFailed,
+			now,
+			notificationmodel.RecordStatusSending,
+			sendingStaleBefore,
+		).
+		Updates(map[string]any{
+			"status":        notificationmodel.RecordStatusSending,
+			"next_retry_at": nil,
+			"updated_at":    now,
+		})
+	if result.Error != nil {
+		return false, result.Error
+	}
+	return result.RowsAffected == 1, nil
+}