Performance testing is one of the important non functional testing methodology. Performance testing is as important as functional testing, even when an application is functionally working as expected if it is not performant enough there are high chances of having a negative impact in the business. so it is the duty of every product owner to make sure that the application is working as expected functionally and non functionally.
There are various performance testing tools available in the market such as load Runner, Neo Load, Jmeter etc.. Load Runner was one of the major market player for a long period of time. However, industries around the globe started searching for value added open source performance testing tools. Jmeter has become a natural fit for this position due to its ease of use and frequent community support. JMeter is now one of the key player among the performance testing tools even though it doesn’t support a lot of protocols as LoadRunner do.
Activities that are commonly performed in Jmeter
- Record the business journey to test.
- Customize the recorded script to feed dynamic parameter values etc. Example: Dynamic user details.
- Provide transaction names for each business clicks such as launch application, Login to application etc to measure the response times for each of these business clicks.
- Provide virtual user load as per the volume expected by the business.
- Execute different type of performance testing.
some of the types of performance tests that can be carried out with JMeter:
- Load Testing: Simulate load in the system for a certain duration to check the response time of key business transactions under the expected load.
- Stress Testing: Stress the system by increasing the load further to check how application is behaving under excess load.
- Endurance Testing: This is to manly check how the application is behaving when executed for a prolonged time period under a given load. Mainly this test is done to check memory leaks.
- Failover & Recovery Testing : Failover testing for an example is done by bringing down one of the server in the cluster if there are 2 servers and to asses if the other server is able to handle the application with out any response time impact.