What constant throughput and linear respone time tells about the system?

Question

What does it mean when throughput in the system is constant and only response time changes (linearly)?

Does it mean that I have to have a bottleneck?

Answer 1

If, at increasing load, throughput does not change, this generally means that your system is running at full capacity. If, on the other hand, a load increase brings added throughput (even if it costs some latency), your system is not running at maximum capacity.

A good example are disk IOPS: at low load (ie: QD1), they often are much lower than what you can expect at high load (ie: QD16).