Can i set ramp up period 0 in JMeter? [closed]

问题:

I am doing performance load testing in a web based application.

I learnt from JMeter Tutorial that how can I set the Ramp Up period, eg:

If 10 threads are used, and the ramp-up period is 100 seconds, then JMeter will take 100 seconds to get all 10 threads up and running. Each thread will start 10 (100/10) seconds after the previous thread was begun.

10 users, 100 seconds ramp-up - start with 1 user, each 10 seconds 1 user added

but I have one doubt can we set Ramp Up period 0 ? if yes,then How result will be affected by 0 ramp up?

回答1:

Ramp up of zero (which is equivalent to 1 sec rampup) is rarely useful, except for large sites that promote release dates, ie when a new product is released at a schedule such as concert tickets or pc games.

You can use it to prove the site would cope with a sudden large influx of requests.

Likely you might also find use for a stepping ramp up, which is available from jmeter-plugins.

As others say, you are trying to simulate normal usage, so if normal usage includes sudden heavy load, keep rampup low.

However, bear in mind that with low rampup and many threads, you may be limited by local resources, so your results may be a measurement of client capability rather than server.

Look at distributed load testing, to simulate even greater loads than a single client can achieve.

回答2:

As stated on the JMeter site the reason for the ramp-up period is:

Ramp-up needs to be long enough to avoid too large a work-load at the start of a test, and short enough that the last threads start running before the first ones finish (unless one wants that to happen).

As an example, say you have a simple test plan that does something like:

1. User logs in.
2. User adds some content.
3. User reads some content.
4. User logs out.

Lets say you set it up to use 1000 threads and and have a ramp-up period of 0 then when you start your test run it will make 1000 log in requests at the exact same moment (give or take a bit).

Needless to say that this is not a realistic test. Unless you are experiencing some kind of DDoS you would never expect peaks in usage like this.

Your goal is to simulate a usage profile that is as close as possible to how your system is going to be used in the real world and for that to happen it is likely that you wish to evenly distribute usage across all of the actions in your plan. For this you use the ramp-up period. An easy rule of thumb is to set the ramp-up period to be the same as the amount of time you expect a single run through of the plan to take.