How does the Constant Throughput Timer control request rate?

The Constant Throughput Timer calculates a delay to add before each sample so that, averaged across the threads it applies to, requests are sent at a target rate expressed in samples per minute, rather than controlling response time or thread count directly.

It works by dynamically adjusting the pause length based on recent actual throughput - if the test is running faster than the target rate, it inserts a longer pause; if slower, a shorter one - to steer the average toward the configured target over time rather than enforcing an exact, rigid interval on every single request.

A key limitation is that it can only slow requests down toward a target rate, not speed them up beyond what the configured thread count and think times can actually achieve - if there simply aren't enough threads generating requests fast enough on their own, the timer can't manufacture additional throughput to hit a target that's higher than the test's natural capacity.