The application features a trap rate limiter. The trap rate limiter limits the rate at which QoS traps are sent. The rate limiter protects against overloading the trap manager with bursts of traps when a single event causes multiple RTP sessions to terminate simultaneously.

The trap rate limiter uses a token bucket scheme, in which traps are sent only if there are tokens in a virtual bucket. Tokens are added to the bucket every 'token interval,' which sets the maximum long term trap rate. Each time a trap is sent, the number of tokens in the bucket decrements. The 'bucket size' is the maximum number of tokens that the bucket can hold. The bucket size limits the trap burst size.