Breaking things before they break you: Load-testing livestream #2 recap

Join us for a spooky load-testing adventure as we explore how to break things before they break you! In this livestream, we simulated massive real-world traffic scenarios on an e-commerce application, showcasing dynamic scaling and real-time observability.

By Thomas di Luccio, on Nov 04, 2024

We recently hosted a second Upsun livestream that was anything but ordinary—we embraced the Halloween spirit by literally putting our application to the test.

With the energy high, Greg Qualls and I set out to break things live, running through a spooky but insightful load-testing session. The recording is now available to catch all the action, insight, and surprises along the way:

What we covered

Load-testing before peak times

The session’s focus was straightforward: let’s break things before they break us. With Black Friday and other peak events looming, we explored how to use Locust—an open-source tool—to simulate real e-commerce traffic scenarios. From casual browsers to quick buyers, we aimed to capture the diversity of customer behavior.

The setup

We set up the Shopware application with Python-driven Locust, creating realistic user scenarios beyond hammering the site with traffic. The goal? To pinpoint breaking points, discover bottlenecks, and see where improvements could be made in a real-world setup before it’s too late.

Scaling resources on the fly

Throughout the live stream, we showcased the Upsun platform’s flexibility to dynamically add resources as traffic scaled—all in real-time. This means no need to alert DevOps teams to scramble, no downtime worries, just pure agility. We’re all about ensuring you can handle even the spookiest traffic levels without breaking a sweat.

Observability in real-time

An essential part of load testing is observability. We talked about how Blackfire let us monitor what’s happening under the hood during heavy loads—where the application’s performance starts to degrade, and exactly where bottlenecks appear. This kind of visibility is crucial not just for testing but for communicating with stakeholders about scaling and performance limits.

Why you should care

Load testing is not just about sending a lot of traffic to your application. It’s about reproducing real scenarios, learning their consequences, and knowing where you’re vulnerable before peak times hit.

It’s about being proactive instead of reactive. Whether preparing for Black Friday or just wanting to sleep better at night knowing your app is resilient, load testing is worth the effort.

If you’re curious to see how we pushed the limits of our application, increased traffic loads, and tested scaling solutions—watch the full recording. And if you have questions or need help with load testing your own application, feel free to reach out in our Discord server. Join the discussion and let us know your experiences with load testing! We’re all in this spooky testing adventure together!

To better observability and beyond

Thomas di Luccio

Thomas is a Developer Relations Engineer at Platform.sh for Blackfire.io. He likes nothing more than understanding the users' needs and helping them find practical and empowering solutions. He’ll support you as a day-to-day user.