Laracon EU: One Month In, Three Regions Deep

I joined Laravel as a contractor in October 2024, then went full-time in January. By the time Laracon EU arrived in February, I'd been on the team for a few months but full-time for barely a few weeks. That isn't a lot of time to get comfortable. But the keynote was happening regardless, and the infrastructure underneath the demo had to work.

Three regions for the keynote

My job in those first months was standing up the multi-region infrastructure that Laravel Cloud would demo on stage. Three full regions, each independently operational, each capable of running real workloads. Production-grade, demoed live in front of a thousand developers.

We had been stress testing for months. Load tests, failure injection, region failovers. The system was solid. I knew it was solid. But knowing it works and trusting it during a live demo are two different things.

Backstage

During the keynote, I was backstage with my laptop open next to Joe Dixon. Our job was simple: watch the dashboards while Taylor ran the demo. Dashboards, logs, metrics. If something went wrong, we would see it first.

Taylor is on stage. The demo is going well. He walks through the features, the audience is engaged, the infrastructure is humming along quietly in the background. Then he clicks a button.

Nothing happens.

At that exact moment, my console goes absolutely red. Errors everywhere. My heart stopped. Joe and I are staring at the screen, trying to process what we are seeing. Had the platform crashed? It couldn't have. We had stress tested this for months. Every failure scenario we could imagine, we had run. This wasn't supposed to happen.

Taylor, being Taylor, stays calm on stage. He scrolls down. And there it is: a validation error. The form had a validation message that wasn't visible because the screen was zoomed in for the projector. He had missed a required field. The platform was fine. It was doing exactly what it was supposed to do: rejecting invalid input.

And my red console? An early access customer had chosen that exact moment to deploy a broken application. The errors I was seeing hadn'thing to do with the keynote demo. The infrastructure was handling it correctly, isolating the bad deployment, showing the errors for that specific app. The platform was working perfectly. It just happened to look terrifying at the worst possible moment.

I aged about five years in those ten seconds.

The hackathon

Around the conference, the Cloud team ran a hackathon. The entire team in one room, heads down, building features and fixing things together. Being in the same room made everything faster.

I'd been talking to these people on Slack for months. Sitting next to them, debugging the same system on adjacent laptops, is a different experience entirely. Different experience entirely.

The team

This was the first time I met most of my colleagues in person. I knew their voices, their code, their opinions in pull request comments. Now I knew their faces and how they ordered coffee, and that changed things more than I expected.

The takeaway

The infrastructure held. That's the story. Not the scare. The fact that when a customer deployed a broken app mid-keynote, the platform just handled it. The red console was monitoring showing us a contained problem, not a crash.

Good luck telling your heart rate that while you're backstage watching it happen live.

The next months would bring talks at meetups across Europe and eventually Laracon US. But Laracon EU was where it started to feel real. And where I learned that the worst moment to have a customer deploy a broken app is when a thousand people are watching your platform on a projector.