We’re in the thick of the heaviest traffic season for many sites and apps. Autumn through the end of the year often brings major traffic spikes, be they from Singles Day, Black Friday, or other seasonal retail peaks. But no matter when your site’s traffic peaks, there’s one big pitfall you need to avoid:
Don’t stop testing after your peak traffic period.
We’ve learned that healthy sites mean consistent testing. If you want to keep your site healthy, help it with continuous performance testing. We’re using a couple of similar words: both consistent and continuous. They’re similar, but not the same.
It’s essential to be consistent with your performance testing if you want to keep your site healthy. We’ve seen it many times: it’s easy to stumble into the pitfall of testing only before peak traffic. While that’s better than no testing, it means a lot of work during one part of the year. It means long days and nights. And it means you’re making things too hard on yourself and your team.
Instead, we recommend consistent performance testing. By “consistent,” we mean regularly and on a schedule. You may not feel like you are ready to integrate it into your regular builds (see below). Regardless, you need to consistently test performance just like you consistently perform unit tests.
Consistency helps you better distribute the burden of optimizing site performance across the year, not in one rush period. When you’re piling all of your performance optimization in one period, you’re inviting shortcuts. You’re not able to spend the time optimizing each piece of performance, instead having to settle for “good enough.”
Instead, when you consistently measure performance, you know when issues arise soon after they arise. For example, perhaps a content provider, internal or external, has just provided you with a big product listing update. Hypothetically, let’s say they forgot to optimize images, or forgot to use your content delivery network (CDN). You’d catch that early on and be able to fix it as it happens, not months later.
Or, for another example, perhaps an API you use has an update and that update has a drastically negative effect on your site performance. Catch it early and you have time to address it. Catch it at the last minute and you have fewer options, especially with such an external dependency.
We also advise continuous performance testing for optimal site health. We choose and use this word because we suggest you integrate performance testing in your regular build cycles using your continuous integration (CI) or continuous delivery (CD) tools.
You’ll find that performance tests conducted as part of your CI/CD continuous practice automates the consistency you want. You don’t have to remember to run consistent tests - they can happen automatically with daily builds. You can set performance thresholds and flags to make you aware of issues so you can fix them early.
When you perform your performance tests consistently and continuously, you’re able to better keep your site healthy. It’s good practice for managing your team and being ready no matter what traffic your site experiences.
We’re here to help if you have any other questions: just ask!