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/server/routes/gateway.go

@@ -18,8 +18,11 @@ func RegisterGatewayRoutes(
 	subscriptionService *service.SubscriptionService,
 	cfg *config.Config,
 ) {
+	bodyLimit := middleware.RequestBodyLimit(cfg.Gateway.MaxBodySize)
+
 	// API网关（Claude API兼容）
 	gateway := r.Group("/v1")
+	gateway.Use(bodyLimit)
 	gateway.Use(gin.HandlerFunc(apiKeyAuth))
 	{
 		gateway.POST("/messages", h.Gateway.Messages)
@@ -32,6 +35,7 @@ func RegisterGatewayRoutes(
 
 	// Gemini 原生 API 兼容层（Gemini SDK/CLI 直连）
 	gemini := r.Group("/v1beta")
+	gemini.Use(bodyLimit)
 	gemini.Use(middleware.ApiKeyAuthWithSubscriptionGoogle(apiKeyService, subscriptionService, cfg))
 	{
 		gemini.GET("/models", h.Gateway.GeminiV1BetaListModels)
@@ -41,10 +45,11 @@ func RegisterGatewayRoutes(
 	}
 
 	// OpenAI Responses API（不带v1前缀的别名）
-	r.POST("/responses", gin.HandlerFunc(apiKeyAuth), h.OpenAIGateway.Responses)
+	r.POST("/responses", bodyLimit, gin.HandlerFunc(apiKeyAuth), h.OpenAIGateway.Responses)
 
 	// Antigravity 专用路由（仅使用 antigravity 账户，不混合调度）
 	antigravityV1 := r.Group("/antigravity/v1")
+	antigravityV1.Use(bodyLimit)
 	antigravityV1.Use(middleware.ForcePlatform(service.PlatformAntigravity))
 	antigravityV1.Use(gin.HandlerFunc(apiKeyAuth))
 	{
@@ -55,6 +60,7 @@ func RegisterGatewayRoutes(
 	}
 
 	antigravityV1Beta := r.Group("/antigravity/v1beta")
+	antigravityV1Beta.Use(bodyLimit)
 	antigravityV1Beta.Use(middleware.ForcePlatform(service.PlatformAntigravity))
 	antigravityV1Beta.Use(middleware.ApiKeyAuthWithSubscriptionGoogle(apiKeyService, subscriptionService, cfg))
 	{