Hi, We run a farm of radius proxies for many different customers (wholesale of services), our devices send requests to the proxies, which in turn send them to the proper backend systems. Number of requests doesn't usually exceed 40-60 a sec, however if one of the client devices gets restarted (or simply crashes) the limit of requests (8192) gets depleted very quickly and proxies start dropping requests from all of the clients, sometimes the backend system of given customer gets overwhelmed which causes additional problems. Is there a way of setting the number of requests on per-client basis, so a single device can't use up whole pool? Of course I can increase the total number of request even further but that only postpones the problem. kind regards pshem