0
votes
  1. I write a JMeter test and use 1000 threads, and get a throughput of 330 requests per second. What was the average response time?
  2. same test in number 2 and I use 100 threads and get a throughput of 330 requests per second. What was the average response time?

I think it has to do with little law, but I have no idea how to solve it? Any help, thanks.

1
You use the same calculation that you would use for any other average, like the average cost of your phone bill per month. There's nothing different involved here. It's basic math. - Ken White

1 Answers

0
votes

We don't know, in order to determine the average response time we need to know your test duration

JMeter calculates average response time as arithmetic mean or the all response times for individual samplers, it can be observed in i.e. Aggregate Report listener.

Also the fact you have the same throughput for 100 and 1000 users looks utterly suspicious, for well-behaved application you should get 10x times more throughput for 1000 users than for 100 users.

The reasons could be in:

  1. Your application cannot handle more than 330 requests per second which indicates a performance bottleneck
  2. JMeter cannot produce more than 330 requests per second, make sure to follow JMeter Best Practices or consider Distributed Testing if your load generator hardware specifications are too low to produce the required load.