perf(后端): 完成性能优化与连接池配置

新增 DB/Redis 连接池配置与校验，并补充单测

网关请求体大小限制与 413 处理

HTTP/req 客户端池化并调整上游连接池默认值

并发槽位改为 ZSET+Lua 与指数退避

用量统计改 SQL 聚合并新增索引迁移

计费缓存写入改工作池并补测试/基准

测试: 在 backend/ 下运行 go test ./...

backend/internal/service/openai_gateway_service.go

@@ -768,20 +768,12 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec
 	if isSubscriptionBilling {
 		if cost.TotalCost > 0 {
 			_ = s.userSubRepo.IncrementUsage(ctx, subscription.ID, cost.TotalCost)
-			go func() {
-				cacheCtx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-				defer cancel()
-				_ = s.billingCacheService.UpdateSubscriptionUsage(cacheCtx, user.ID, *apiKey.GroupID, cost.TotalCost)
-			}()
+			s.billingCacheService.QueueUpdateSubscriptionUsage(user.ID, *apiKey.GroupID, cost.TotalCost)
 		}
 	} else {
 		if cost.ActualCost > 0 {
 			_ = s.userRepo.DeductBalance(ctx, user.ID, cost.ActualCost)
-			go func() {
-				cacheCtx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-				defer cancel()
-				_ = s.billingCacheService.DeductBalanceCache(cacheCtx, user.ID, cost.ActualCost)
-			}()
+			s.billingCacheService.QueueDeductBalance(user.ID, cost.ActualCost)
 		}
 	}