I built a tool inspired by eksctl, but on top of Crossplane. It operates on versioned, validated configuration data, not directly on the live state. The experience of using tools can be so much better than dealing with config as code. confighub.com/blog/buildin...
Soo what's the hottest s3 compatible storage to selfhost right now? We tried garage, since it seemed interesting, but they have no official helmchart and we're seeing problems with a single node/pod deployment. Next on our list would be seaweed. Minio is completely off the rails in slop I've […]
The certificate rotation was one line of code. The approval process was two months. cert-manager solves the easy part—here's how to design away the operational debt that eats the rest of your quarter.
AI agents now write AND run code you never reviewed. Treating that as a "trusted product" is a security gap—and plain container isolation won't save you. Here's why Kata Containers' microVM boundary is becoming table stakes for agent sandboxes.
Live migrate a VM across Kubernetes nodes and the TCP connection survives—no dropped packets. The secret? Calico keeps the pod IP pinned while BGP reconverges the route in real time. Here's how it works under the hood.
Green CI doesn't prove the image running in prod is the one you built. Tags are mutable—anyone can overwrite them. Here's how Sigstore + Kyverno reject unsigned images on K8s, straight from your GitOps flow.
Live broadcasters send every camera feed twice over separate paths and just drop whichever packet loses the race. "Wasteful" redundancy is exactly why your K8s multi-AZ setup survives failures. Same philosophy, different scale.
Your AI-generated YAML passes kubectl apply and "works" — while quietly burning 69% of your cloud bill on overprovisioned requests/limits. Here's what to check before you ship to prod.