Understanding Ramp-up Setting in JMeter

Mary ZhengDecember 20th, 2024Last Updated: December 16th, 2024

0 36 3 minutes read

1. Introduction

Apache JMeter is an open-source software designed for performance testing and load testing of applications. The load testing is completed by simulating virtual users via threads, sending multiple types of requests to the target server, and measuring its performance under varying loads. Understanding the number of threads and Ramp-up periods is important when configuring the thread group.

2. JMeter Concept Overview

JMeter has the following key components:

Test Plan defines the sequence and flow of all test components.
Thread Group represents a group of virtual users (threads) that will execute a series of requests defined in the test plan. Each thread simulates a real user interacting with the application. It configures three properties: Number of threads, Ramp-up period, and Loop count.
Samplers send requests to a server and wait for a response.
Listeners collect the results of test execution and display them in various formats (graphs, tables, logs).
Controllers define the logic flow of the test execution.
Timers add a fixed delay between each request.
Assertions validate the response data to ensure that the server is responding correctly.
Pre-Processors and Post-Processors: Pre-Processors prepare data before a request is sent and Post-Processors handle the response data after a request is executed.
Configuration Elements are the shared configuration settings across different samplers.

3. JMeter Thread Group

In this step, I will explain the Thread Group‘s properties.

jmeter ramp up — Figure 1. JMeter Thread Group

As you seen in Figure 1, there are three thread properties:

Number of Threads (users): each thread will execute the test plan completely independently of other test threads. Multiple threads are used to simulate concurrent connections to your server application. Figure 1 shows the value 10, then it means there will be 10 concurrent connections accessing the server application.
Ramp-up period (seconds) defines the time in seconds to start all threads.
Loop Count defines how many times each thread should repeat the requests. Figure 1 shows the value of 1, then it means only executing once.

Internally, Jmeter calculates the internal period to start the next thread with this formula:

Internal period to start a thread = Ramp-up period / Number of Threads.

If there are 10 threads and the Ramp-up period is 20 seconds, then each thread will be started 2 seconds after the previous thread started.
If there are 30 threads and the Ramp-up period is 120 seconds, then each successive thread will be delayed by 4 seconds.

4. Ramp-up Best Practices

When configuring the Ramp-up period in JMeter, it’s crucial to strike a balance between simulating realistic user behavior and testing the performance of the system under load. The Ramp-up period determines how quickly the threads (virtual users) are started during the test. A proper configuration ensures that the load is applied gradually, allowing the system to handle incoming requests effectively and without sudden spikes that could result in misleading test results. Here are some key notes when configuring the Ramp-up period in JMeter:

Avoid a short Ramp-up period as it can cause all threads to be initiated nearly simultaneously. This could overload the server and give you inaccurate performance results, especially if the application is not designed to handle such quick bursts of load. The system might fail or show degraded performance, not due to a lack of scalability, but because of the unrealistic load applied.
Monitor system behavior during Ramp-up to see how the incoming loads are handled. If the system starts to experience issues during this gradual Ramp-up, it may indicate performance bottlenecks or capacity limits.
Use a Ramp-up period that matches user expectations.

5. Conclusion

JMeter is a powerful tool for simulating real user behavior and load testing. JMeter calculates the interval period to start the thread based on the number of threads and the Ramp-up period. A good rule of thumb is to have a Ramp-up period that is proportional to the number of threads you are using, generally between 1 and 5 minutes. The goal is to simulate real user behavior, avoid sudden traffic spikes, and properly assess the system’s performance under increasing load. Monitoring and adjusting based on the system’s response will ensure more accurate and meaningful test results.