Throttling / rate limiting
A downstream system can often take only so many calls per second, an API quota, a fragile legacy box. Throttling paces the send port: it delivers one message at a time, in arrival order, with a minimum interval between deliveries, so a burst drains at a pace the destination can actually sustain.
In iQBus this is the send port’s Throttle tab, the same for every adapter: switch on Enable Throttle and set Interval (Milliseconds), the minimum time between the start of one delivery and the start of the next. A limit of N calls per second is an interval of 1000 / N: 1000 for one a second, 250 for four. A delivery that takes longer than the interval, or retries, simply pushes the next one back. The full rules are in Retry and Throttle.
Work rarely arrives evenly. A nightly batch or a debatched interchange can dump thousands of messages at once; firing them all immediately would trip a partner’s rate limit or topple a delicate endpoint. A throttle turns that spike into a steady stream: the port queues the backlog, oldest first, and releases it at the configured pace.
Throttling protects the destination; it pairs with the circuit breaker, which protects you from the destination. Slowing down never means dropping, it just spreads the same work over more time: queued messages wait their turn, and one that fails for good ends in error as any delivery would, without blocking the ones behind it. Each port paces only itself, so two ports calling the same server are throttled separately.