fix: move RPM prefetch before routing segment in legacy/mixed paths

Ensures isAccountSchedulableForRPM calls within the routing segment
hit the prefetch cache instead of querying Redis individually.

Author: QTom

backend/internal/service/gateway_service.go

@@ -2607,6 +2607,9 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context,
 		}
 		accountsLoaded = true
 
+		// 提前预取 RPM 计数，确保 routing 段内的 isAccountSchedulableForRPM 调用能命中缓存
+		ctx = s.withRPMPrefetch(ctx, accounts)
+
 		routingSet := make(map[int64]struct{}, len(routingAccountIDs))
 		for _, id := range routingAccountIDs {
 			if id > 0 {
@@ -2822,6 +2825,9 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g
 		}
 		accountsLoaded = true
 
+		// 提前预取 RPM 计数，确保 routing 段内的 isAccountSchedulableForRPM 调用能命中缓存
+		ctx = s.withRPMPrefetch(ctx, accounts)
+
 		routingSet := make(map[int64]struct{}, len(routingAccountIDs))
 		for _, id := range routingAccountIDs {
 			if id > 0 {