监控

internal/polling/queue_manager.go

@@ -186,3 +186,66 @@ func (m *PollingQueueManager) GetTotalQueueDepth(ctx context.Context, taskType s
 	}
 	return total, firstErr
 }
+
+// GetTotalDueCount 获取指定任务类型所有分片中已到期的队列数量。
+// 只统计 score <= now 的成员，用于后台监控展示真实待调度积压。
+func (m *PollingQueueManager) GetTotalDueCount(ctx context.Context, taskType string, now time.Time) (int64, error) {
+	var total int64
+	var firstErr error
+	for i := 0; i < m.shardCount; i++ {
+		key := constants.RedisPollingShardQueueKey(i, taskType)
+		count, err := m.redis.ZCount(ctx, key, "-inf", fmt.Sprintf("%d", now.Unix())).Result()
+		if err != nil {
+			m.logger.Warn("获取分片到期数量失败",
+				zap.Int("shard_id", i),
+				zap.String("task_type", taskType),
+				zap.Error(err))
+			if firstErr == nil {
+				firstErr = err
+			}
+			continue
+		}
+		total += count
+	}
+	return total, firstErr
+}
+
+// GetAverageWaitTime 获取指定任务类型所有分片最早若干任务的平均等待秒数。
+// 每个分片最多取 samplePerShard 条，避免监控接口扫描完整队列。
+func (m *PollingQueueManager) GetAverageWaitTime(ctx context.Context, taskType string, now time.Time, samplePerShard int64) (float64, error) {
+	if samplePerShard <= 0 {
+		samplePerShard = 10
+	}
+
+	var totalWait float64
+	var sampleCount int64
+	var firstErr error
+	nowUnix := now.Unix()
+
+	for i := 0; i < m.shardCount; i++ {
+		key := constants.RedisPollingShardQueueKey(i, taskType)
+		earliest, err := m.redis.ZRangeWithScores(ctx, key, 0, samplePerShard-1).Result()
+		if err != nil {
+			m.logger.Warn("获取分片最早任务失败",
+				zap.Int("shard_id", i),
+				zap.String("task_type", taskType),
+				zap.Error(err))
+			if firstErr == nil {
+				firstErr = err
+			}
+			continue
+		}
+		for _, z := range earliest {
+			waitTime := float64(nowUnix) - z.Score
+			if waitTime > 0 {
+				totalWait += waitTime
+			}
+			sampleCount++
+		}
+	}
+
+	if sampleCount == 0 {
+		return 0, firstErr
+	}
+	return totalWait / float64(sampleCount), firstErr
+}