Clever Geek Handbook

This is the best performance testing tool when I don't know the behavior of the client application. Since I will have a number of web clients

My organization wants to do performance testing externally. Performance testing will be done for web applications only. Can you please advise me which tool is the best to choose as a performance testing tool. Right now, there is no clue about the behavior of the client web application. The number of clients can be avoided. What is the minimum configuration I should be looking for hardware?

Can I get 100 percent accurate results using the open source tool on the market? I am really amazed by these inner doubts. I am new to performance testing. Your answers are very helpful

+3


source to share


3 answers


I am new to performance testing.

Then you need to have a mentor to work with for a period of time. The tool chosen represents about 5-15% of the suite of performance test solutions.

Write down your requirements for reporting, analysis, your team's skills, applications and hosts you need to monitor, in the environment in which your tool should run. Use these requirements as a lens with which you can view your settings.



In terms of accuracy, I would suggest that you take the following into account as part of your testing efforts: technically evaluate the technique. Look at the load from a single live user in your field in terms of IP connections, HTTP sessions, etc., then using the same tools, see how your tool generates its load. You will find some differences. Some tools, which are quite popular, will use fewer sessions and piggyback requests from multiple users over an open connection. This will reduce the number of connections compared to the natural population and make it harder to work on bottlenecks in the session. I will not identify the tool / tools that are so inclined that will be for your research. You will want to do this assessment for one,five and ten users to see how your load is shaped and the structural differences between tools.

Large commercial vendors have smaller virtual custom versions of their tools for free. You can use these versions to assess your requirements and your technical assessment of compliance with true user behavior.

+4


source


If you are looking for a free open source benchmarking tool, I would recommend skipping to the next one:

They are more or less equal in their load testing capabilities when it comes to web applications, however they differ from each other, i.e.



  • You need to know Python if you are using Grinder
  • Gatling offers an easy-to-read DSL for developing test scripts, and customization is done with Scala
  • Tsung can generate more load when using the same hardware, but it does not provide any GUI and only works with Linux and Unix.
  • JMeter development is done through Tree of Components in a GUI, and there is a DSL if you need code and no GUI. Customization of scripts can be done with any language that provides the JSR223 bridge (Groovy, Scala, Java, JRuby ...).

So, if you are new to performance testing, I would recommend Apache JMeter. However, you can check:

+4


source


jMeter is a good answer here as it is free, you can easily write some scripts if you have never worked with such tools and if you are thinking about loading tests for 1k threads (users) or still it is quite easy to do. you just need to get remote tested (link here ).

And something about tips and tricks , this is really helpful when you go through some tutorials, write some tests, and then you will see that this is not what you wanted.

+2


source






More articles:


All Articles