The Hidden System Behind 1 Billion Build Manifests

Michael Miller · 2026-09-08

In just six months, Chainguard doubled its output from 500 million to over 1 billion container build manifests. That's a lot of zeros, right? But the real story isn't the number. It's the machine that made it possible. We also blew past 3,000 unique container images and 675,000 image versions. Those are the headline stats, but they don't tell you why we had to tear down our old approach and rebuild from scratch. Let me pull back the curtain. ### The Numbers Are Just the Surface When you're shipping that many manifests, you can't rely on manual processes or duct-taped scripts. You need a system that scales automatically, catches errors before they cascade, and keeps every image consistent. That's what we built. But getting there meant admitting our previous setup couldn't keep up. We had to rethink everything: how we store images, how we version them, how we push updates without breaking downstream users. It was less about hitting a billion and more about building something that could handle the next billion. ### Why We Had to Fundamentally Change Our old pipeline worked fine at 500 million. But as we grew, bottlenecks appeared. Builds took longer. Storage costs ballooned. And debugging a single failed image felt like finding a needle in a haystack. So we redesigned the core. We moved to a more distributed architecture, automated validation at every step, and introduced smarter caching. The result? Faster builds, fewer errors, and a catalog that's easier to navigate. Here's what that looks like in practice: - **Automated validation** catches issues before they reach production. - **Distributed builds** spread the load across multiple machines. - **Intelligent caching** reduces redundant work and speeds up delivery. - **Unified versioning** makes it simple to track changes across thousands of images. ### What This Means for You If you're managing containers at scale, you know the pain of inconsistent images and slow builds. Our journey shows that you don't have to settle for a fragile system. With the right architecture, you can scale without sacrificing reliability. And it's not just about raw numbers. It's about giving developers confidence that every image they pull is secure, up-to-date, and exactly what they expect. That's the real win. So next time you see a billion of anything, remember: the number is just a symptom. The system behind it is what matters.