Foundations that hold, even when your cloud provider doesn't
What happens to your checkout when your cloud provider has a bad day? For most retailers, the answer is: the same thing that happens to everyone else on that provider. We decided that wasn't good enough. In this Insight, Mark Gerrits explains how he built a unified commerce platform that keeps running when any single provider fails. Identical frontends are served from two independent CDNs, backend clusters are spread across clouds and regions, and traffic fails over in under a minute, without customers noticing. You'll learn: Why single-provider setups quietly cost you uptime, negotiating leverage, and flexibility | How multi-cloud can stay simple by building the same thing everywhere | How data residency and GDPR turn redundancy into a compliance advantage | How to rebuild your foundation without closing the store. Plus three questions to ask about every critical dependency in your own stack.
Why we stopped depending on any single cloud provider, and what it taught us about building things that last.
The teams that build infrastructure on a single provider are making a strategic concession, whether they know it or not. They are handing uptime, negotiating leverage, and the ability to change to a company whose interests are not theirs. Those three things turn out to be the same decision, and recovering them starts with the same place: making sure no single vendor can take down everything you have built. Every application you run stands on something: a pile of services, providers, and quiet assumptions that nobody thinks about, right up until one of them moves. The question most teams avoid is not whether that will happen. It is whether they built something that survives it when it does.
For a long time our answer was simpler than we wanted to admit. One provider handled our hosting, our compute, our caching, our TLS, and our routing. It worked well. That was exactly the problem. When something works, you stop questioning it, and before long that one provider is holding up everything you have built. An outage is no longer an inconvenience. A pricing change is no longer a line item. A policy shift is no longer someone else's problem. Each of them becomes a crack in the foundation, and you find out the hard way that you were renting the ground you built on.
Unlock this content
Join the NERD. Community to access articles, badges, interviews, and exclusive community insights.