Dora Noda
Software Engineer
2304 posts · View all authors
IPv6-Only Worker Nodes on Hetzner: What Dropping the €0.50 IPv4 Line Item Saves a 10–50 Node Fleet, and What Breaks First
Hetzner bills every primary IPv4 at €0.50/month while IPv6 is free. The real savings math for a 10-50 node CAPH fleet after the June 2026 repricing, why GitHub and ghcr.io break first on IPv6-only nodes, and the four bridges — public NAT64, Jool, a NAT gateway, or a registry cache.
Kubernetes Just Buried Its Own Dashboard: Why Headlamp Is Now the Default Operator UI for a CAPI-Managed Fleet
The Kubernetes Dashboard is archived and the project's own blog now points operators to Headlamp. What the June 2026 migration guide and the new Cluster API plugin mean for a CAPI-managed fleet: which plugin features cover which operator surfaces, and when embedding Headlamp beats building a bespoke dashboard.
Why Self-Hosted PaaS Secrets Managers Still Default to Plaintext Env Vars: A Survey of Coolify, Dokploy, and CapRover
Coolify, Dokploy, and CapRover all hand your production credentials to anyone with dashboard access — a feature-by-feature survey of how each stores secrets, why plaintext env vars are an architectural default, and what a Sealed Secrets or External Secrets Operator pattern on Kubernetes buys instead.
The Tailscale Kubernetes Operator Exposes kube-apiserver Without Touching the Public Internet
84% of internet-observable Kubernetes API servers answer to the public internet. The Tailscale operator's API server proxy replaces the IP allowlist with identity-authenticated tailnet access — here is what changes for a self-hosted fleet, what it costs, and what Headscale users shouldn't assume.
Vercel's Four Repricings Since 2024: What the Pattern Predicts for Your Next Bill Shock — and What a Fixed-Price Box Actually Fixes
Vercel re-modeled its billing four times in eighteen months — granular meters, Fluid Compute, Active CPU, then credits. Here is the full timeline, what the 3-10 month cadence means for your budget, and a worked comparison against a flat-rate Hetzner box where egress variance is zero.
xAI Open-Sourced Grok Build — But You Can't Send a Patch: What Apache-2.0 Without Open Contributions Actually Grants You
xAI published the 845,000-line Rust harness behind Grok Build under Apache-2.0 — then refused all external pull requests. A concrete breakdown of what a permissive license without open contributions actually grants a team: audit and fork rights, no roadmap influence, and a security-patch treadmill.
Cluster API v1.12's In-Place Updates and Chained Upgrades: What Changes for a Platform Team Running CAPH/CAPD in Production
Cluster API v1.12 shipped in-place updates and chained upgrades. For a 3-control-plane, 20-worker Hetzner fleet jumping three Kubernetes minors, that turns 29 replace-and-drain cycles under best pre-v1.12 practice into as few as zero — and three hand-sequenced hops into one declared version. The before/after runbook, with the CAPD rehearsal loop.
Flagger vs Argo Rollouts in 2026: Which Progressive-Delivery Operator Should Power a Git-Push PaaS's Canary Rollouts?
Flagger and Argo Rollouts shipped releases three days apart in July 2026, but they make opposite bets: wrap the Deployment or replace it, headless reconciler or human dashboard. A build-vs-choose comparison of each operator's stack assumptions, with a concrete verdict for a git-push PaaS on a Cluster API fleet.
ArgoCD 3.3 vs Flux 2.0 in 2026: Which GitOps Reconciler Fits a Cluster-API-Managed Fleet's Own Deploy Pipeline
Argo CD has 2.8x Flux's GitHub stars, but a self-hosted PaaS on Cluster API needs a reconciler shape, not a deploy dashboard. A four-axis head-to-head — tenancy, fleet rollout, footprint, security — with a concrete verdict for platform builders.