Cilium Becomes the Default Kubernetes CNI Everywhere but AWS: What eBPF Networking Buys a Self-Hosted PaaS on Hetzner
GKE and AKS now run Cilium by default while EKS holds out — here's what eBPF service routing, Hubble flow observability, and identity-based network policy actually buy a self-hosted PaaS on Hetzner bare metal, with the benchmarks, memory costs, and a decision table.
Blue-Green Deploys on Cluster API: Flipping a Tenant's Traffic Between Two MachineDeployments Without Touching the Load Balancer
A rolling MachineDeployment update isn't blue-green. Here's the actual Cluster API primitive for an atomic, instantly-reversible traffic cutover — two MachineDeployments, one Service selector flip, and the in-flight-request handling most implementations skip.
The Buildpack Escape Hatch: Why 'No Dockerfile' Breaks for About 1 in 5 Repos
Buildpack pitches promise you'll never write a Dockerfile — until the detect phase rejects your repo with no editable file to fix. A worked estimate of how often that happens, and the three-tier escape hatch a git-push platform should ship instead of just falling back to raw Dockerfiles.
Buildpacks vs Dockerfiles for a Git-Push PaaS: The Real Build-Time, Image-Size, and Leaked-Detail Numbers
Buildpacks promise 'no Dockerfile needed,' Dockerfiles promise full control. A real, sourced comparison of build time, image size, and how much implementation detail each one pushes back onto the developer.
Show HN's Cygnus: What a Serverless-Runtime Framing Actually Changes (and What It Doesn't) Versus a Container-Per-App PaaS
Cygnus revives an idle app from zero in about 30 milliseconds using kernel-sandboxed 'cages' instead of containers. Here's exactly what that architecture buys over Coolify and Dokploy's always-on containers — and why it still hits the same one-box ceiling the rest of the category has.
Your Git-Push PaaS Assumes GitHub. Here's What Wiring It to Forgejo + Woodpecker Actually Takes
A git-push PaaS built GitHub-first assumes a webhook, a payload shape, and a status write-back that don't just carry over to a self-hosted Forgejo + Woodpecker CI stack. Here's the concrete four-piece build-out — webhook receiver, deploy trigger, commit-status write-back, OAuth app — that makes it a first-class integration instead of a race condition.
GitHub Wobbled Three Separate Times in Eight Days: What Owning No Runner Capacity Costs
Three unrelated GitHub subsystems failed independently within eight days in July 2026. A concrete accounting of what broke, why that's a different risk shape than one outage repeating, and what actually changes when the build step and deploy target run on owned infrastructure instead.
The 2026 'PaaS-First' Reversal: What Hacker News Gets Wrong About Owning Your Kubernetes Cluster
Hacker News sentiment flipped from 'you don't need Kubernetes' in 2024 to pitching Kubernetes-without-a-platform-team products in 2026 — but the PaaS-first argument conflates hand-operated Kubernetes with vendor lock-in PaaS, missing the option that actually removes the toil: Cluster API-based ownership.
Java 26 Will Never Be Paketo's Default — What That Says About the Treadmill a Buildpacks Platform Runs Forever
Paketo added Java 26 support in July 2026, but under its own RFC #0014 math, Java 26 will never be the default — and gets dropped the moment Java 27 ships. A concrete accounting of the version-tracking treadmill a buildpacks platform runs forever, and what a Dockerfile-only platform pushes onto app authors instead.