What if you could take one of the biggest React SSR platforms in the Netherlands, swap its runtime, and make it faster, leaner, and more predictable—all in one week?

That’s exactly what we did. Node.js was hitting its limits: high latency, memory spikes, and pods scaling aggressively during traffic peaks. Every engineer felt the pain—SSR performance was unpredictable under load.

We swapped Node for Bun. Same app, same traffic, same production environment—just a different engine.

The results were immediate. Throughput jumped, p99 latency dropped, memory usage nearly halved, and scaling became far smoother. Pods spun up in 7 seconds instead of 30, CPU spikes flattened, and clusters handled bursts with precision.

In this talk, I’ll show how we ran the experiment: dual runtimes in production, safely mirroring traffic, benchmarking performance, and tuning scaling policies. You’ll see real numbers, graphs, and actionable lessons from thousands of concurrent requests.

If you care about React SSR, runtime performance, or want to explore how modern JavaScript engines can transform production workloads, this talk gives you insights and takeaways you can apply immediately.