On 12/7/05, Chetana Anand <[EMAIL PROTECTED]> wrote: > Does any body have an idea on how spike testing can be simulated in > jmeter... Like increasing the no. of concurrent users at different > instants > of time..
You could create, say, three Thread Groups. Suppose your test is 10 minutes long and will start at 12:00. The first thread group has, say, 4 threads and runs from 12:00 to 12:10. The second group has 10 threads and runs 12:04-12:05. The third group has 15 threads and runs 12:07-12:08. Thus, you'll have 10 minutes of normal activity with two bursts of high and then very high activity in the middle. Is that roughly what you wanted to achieve?
