New tool for Load Testing: Load Xen
We’re proud to announce launch of our load testing tool, Load Xen. You can use it to generate high amount of traffic to your application and figure out how your app did. Load Xen collects number of performance metrics during your test which will help you understand how your application performs under stress.
Why we built yet another load testing tool?
While building Load Xen, we continuously reminded ourselves that this is going to be for humans not machines, and that was it, the core idea: simplicity.
Simplicity
If you are new to load testing, you’ll find Load Xen quite easy to use. Within 1 hour you can write customized load tests which you never thought were possible or were too hard to code.
If you are someone who’s fairly familiar with load tests or stress testing, you’ll still save a lot of time in writing tests. Load Xen has its own DSL for creating load scripts, which is nothing fancy, but just simple JavaScript code you’re already familiar with.
Why do I have to write code?
No text fields, text areas or combined fields are as powerful as 10 lines of code. We ran into this decision where we had to choose between a complicated interface involving complex fields or a neat editor, we chose latter. We did because it was much neater, flexible and sexier than traditional approaches. As a result, it offers HUGE flexibility when creating your test scenarios.
Fig. 1, Simple GET request
Fig. 2, randomizing data for requests
I’m not a programmer, what about me?
We have about 7 template scripts for you which can help you get started. You’ll see it’s not that complicated once you create your first test.
Better metrics
When you’re testing your application’s performance a lot of metrics matter:
– how many users are currently online?
– what is the response time?
– are there any timeouts?
– did any of the user received an error?
Such data is generated in large amounts and it probably cannot help you looking at raw data. For a moment think about this:
You are testing an app with 1000 concurrent users for 1 hour, how much records such test will generate every second? Even with 3 simple metrics, that’s 10,800,000 records (1000 * 3600 (seconds) * 3 metrics) and there are several other metrics to be tracked.
We have tried to solve this using simple yet efficient statistics methods, so you don’t have to make sense out of numbers instead you get straight conclusive results. Not only that, you can pinpoint at any second to visualize how (or what) your app did at that instant. This gives you great debugging information should you run into a problem since you will now know the first occurrence of issue.
