Throughput Shaping Timer

Download

Long time JMeter users had to try their tests on and on, finding exact number of threads and timer delays that produce desired number of requests per second (RPS) to server. This is "closed workload" approach, and it has major drawbacks (see here why). And if you changed load generator machine or server response time - you were doomed to do all this stuff again.

Here's the solution - Throughput Shaping Timer! Now you can just set schedule of RPS easily, observing it on preview graph. This schedule can be as various as you want. It gives you the happiness of the "open workload" approach.

Few Important Notes

1. JMeter threads of Thread Groups in scope of the Element will be stopped when RPS schedule finishes.
2. provide enough working threads for your RPS, JMeter timers can only delay threads (and limit RPS). You may pair this plugin with Concurrency Thread Group using Schedule Feedback Function to dynamically maintain thread count required to achieve target RPS.
3. if you're using versions of JMeter lower than 3.3 and if you have RPS that lower at the end of test, make threads to lower also. Оtherwise you'll have a spike in last second.
4. avoid using zero RPS value as start of test, this produce spike also
5. avoid zero RPS during the test, this may lead to nasty effects

How Many Threads I Need To Produce Desired RPS?

Threads pool size can be calculated like RPS * <max response time> / 1000. The more rate desired the more threads you will need. The more response time service have the more threads you will need.

For example, if your service response time may be 2.5sec and target rps is 1230, you have to have 1230 * 2500 / 1000 = 3075 threads.

It is better to have some threads overhead to handle unexpected response time spikes. But beware of too much overhead, it will lead to "final spike".

Example

Example JMX for Throughput Shaping Timer

Shaping Timer setting in example test plan:

Actual RPS provided:

Properties exposed by Component

The element will export the following properties that you can access through __P function or using in JSR223 Test Elements props.get("property name")

Properties are:

• elementName_totalDuration - Total duration as sum of the "Duration,sec" column
• elementName_cntDelayed - Number of current threads waiting, if lower than 1 it means we don't have enough threads to reach RPS
• elementName_cntSent - Number of samples sent since last second tick
• elementName_rps - Current RPS

Special Property Processing

There is a way to specify load pattern from special jmeter property tst-name.load_profile, where tst-name is the name of the Throughput Shaping Timer to define the load pattern on. The load.profile jmeter property is used as a default if the tst-name is not defined. If neither the tst_name.load_profile or default load_profile are defined, the values from GUI are taken Property can be specified either in user.properties file (jmeter.properties also applicable), or from command line like -J "load_profile=...".

Load pattern specified with set of function-like declarations. "Functions" may be of 3 types: const, line and step.

• const(N, T) - hold constant load of N requests per second for T seconds
• line(N, K, T) - linear increase load from N to K RPS during T seconds
• step(N, K, S, T) - generate stepping load from N RPS to K rps, each step height to will be of S RPS, each step duration will be T seconds

Duration value in every function may be specified with shorthand case-insensitive modifiers:

• s - seconds
• m - minutes
• h - hours
• d - days

Duration setting may be combined like 1d11h23m6s.

Example of load profile string: my_timer.load_profile=const(10,10s) line(10,100,1m) step(5,25,5,1h) Example of default load profile string : load_profile=step(5, 50, 5, 2m) const(50, 10m)

Scenario 1:

Outcome:

Scenario 2:

Outcome:

Schedule Feedback Function

There is special JMeter function that enables feeback loop for number of threads. Thread groups like Concurrency Thread Group can use that function to supply additional threads during test runs when requested RPS schedule is not met.

Example function call: ${__tstFeedback(tst-name,1,100,10)} , where "tst-name" is name of Throughput Shaping Timer to integrate with, 1 and 100 are starting threads and max allowed threads, 10 is how many spare threads to keep in thread pool. If spare threads parameter is a float value <1, then it is interpreted as a ratio relative to the current estimate of threads needed. If above 1, spare threads is interpreted as an absolute count.