Techniques for managing transmission rates for users from each of a plurality of call engines (CEs) distributed on a communications network are provided. A list of users with at least one pending call in a corresponding queue from all other CEs is received, wherein each user's queue has a maximum queue depth calculated, in part, using a capped transmission rate in calls per second (CPS) for each user associated with their queue. It is determined, for each user, a total number of CEs having at least one pending call in their corresponding queue. A current CE transmission rate is calculated for each user by dividing the capped transmission rate for each user by the total number of CEs having at least one pending call in their corresponding queue. The current CE transmission rate indicates a rate at which calls in each user's queue are dequeued at each of the plurality of CEs. When a call request is received at a CE from a specific user, a total number of calls currently queued for the specific user in all CEs is determined. A maximum queue depth for the specific user is also determined, the maximum queue depth indicative of the total number of calls the specific user may queue for a predetermined length of time. The call request will be rejected when the total number of calls currently queued exceeds the maximum queue depth.