perf(日志): 降噪优化，将常规成功日志降级为 Debug 级别

- GIN Logger 中间件跳过 /health 和 /setup/status 的请求日志
- UsageCleanup 空闲轮询（no_task）日志降级为 slog.Debug
- Scheduler 常规 rebuild ok 日志降级为 slog.Debug
- DashboardAggregation 常规聚合完成日志降级为 slog.Debug
- TokenRefresh 无刷新活动时周期日志降级为 slog.Debug

生产环境（Info 级别）下自动静默，debug 模式下仍可见。
错误、警告类日志保持原有级别不变。

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/internal/service/dashboard_aggregation_service.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"log"
+	"log/slog"
 	"sync/atomic"
 	"time"
 
@@ -224,11 +225,11 @@ func (s *DashboardAggregationService) runScheduledAggregation() {
 	if updateErr != nil {
 		log.Printf("[DashboardAggregation] 更新水位失败: %v", updateErr)
 	}
-	log.Printf("[DashboardAggregation] 聚合完成 (start=%s end=%s duration=%s watermark_updated=%t)",
-		start.Format(time.RFC3339),
-		now.Format(time.RFC3339),
-		time.Since(jobStart).String(),
-		updateErr == nil,
+	slog.Debug("[DashboardAggregation] 聚合完成",
+		"start", start.Format(time.RFC3339),
+		"end", now.Format(time.RFC3339),
+		"duration", time.Since(jobStart).String(),
+		"watermark_updated", updateErr == nil,
 	)
 
 	s.maybeCleanupRetention(ctx, now)