The goal of performance testing is to get actionable data that will help you optimize your website, app or API performance.
The best way to do that is to create the most realistic load tests possible. We like to dish out tips on how to do that in many different ways. To name a few:
- Use the Load Impact User Scenario Recorder to create your load testing scripts
- Look at your traffic data in Google Analytics when figuring out how many concurrent virtual users you want to test when figuring out your performance baseline
- Run multiple types of tests to see how your app reacts under different kinds of pressure
But this article is going to tackle another topic that we’ve been chatting about with a few different users lately: Considering where you expect traffic to be coming from when configuring your load tests.
Finding where your actual traffic originates
Google Analytics is pretty much the industry standard when it comes to website or application traffic and user data. So, that’s what we’re going to use in this example of how to determine from where you should simulate traffic for your load tests.
From your Google Analytics dashboard, take a look at the left sidebar. Under the header “Audience,” you’ll see a little arrow pointing to the word “Geo.” Click that, and you’ll find a menu item for “location.” Select that, too.
Once you have your desired date range selected, take a look at the breakdown by country. Now that you can see where your traffic is coming from, let’s go back to the Load Impact app and configure where simulated traffic will originate in our test.
How to simulate traffic from different locations
Like any product currently thriving on the Internet, your website’s traffic is likely going to be from a few different parts of the world.
In order to closely mimic the actual traffic you see on a daily basis, you’ll want your load tests to reflect the percentage of traffic that comes from different regions around the world.
When you configure a load test in Load Impact, you’re given the option to select load zones. These are AWS data centers that host our load generators, and we’ve given you eight options to choose from around the world:
Ashburn, Virginia (USA); Sao Paulo, Brazil; Palo Alto, California (USA); Singapore; Portland, Oregon (USA); Sydney, Australia; Dublin, Ireland; Tokyo, Japan
In this example, it looks like the majority of this site’s traffic is coming from Europe, so we’re going to direct most of our simulated traffic from the Dublin, Ireland load generator.
It’s also important to note that you can specify which user scenarios you want to run from each load zone. This is great for when you know different subsegments of your audience are navigating to different pages on your website.
If you look at the example images, this user estimates that about 80 percent of their traffic is coming from Europe.
Thinking about the little details in your load test configuration can mean the difference between getting actionable data from your testing and not really getting a clear picture of your website or application performance.