Our team has identified the responsible server and fixed the problem by balancing traffic rules on the servers