Mattias Åsell

Build a working Kubernetes 1.37 cluster from plain binaries, certificates and systemd units. Every command tested, so you learn what each piece does.

Take an app built with Cursor, Claude Code, Lovable, Bolt or Replit to production in 2026. What production adds, the hosting routes compared, how to pick.

Twelve AI tools for Kubernetes compared, from open source troubleshooters to AI operations platforms and MCP servers, and how each one handles write access.

Flux with a web UI, Akuity, Harness, Octopus, Rancher Fleet, Devtron and Ankra compared, plus when plain Argo CD is still the right call.

Hosted European PaaS or push-to-deploy in your own EU cloud account? Nine Heroku alternatives compared by headquarters, hosting location and fit.

Rancher, OpenShift, Spectro Cloud, Kubermatic, Portainer, Mirantis, Platform9, Azure Arc, Omni and Ankra compared, with who each fits.

Hetzner has no managed Kubernetes. Eight ways to run it in 2026, from kube-hetzner and Talos to Syself, Cloudfleet and Ankra, and what each leaves to you.

Northflank, Qovery, Porter, Ankra, Coolify, Dokploy, Kamal, Render and Railway compared, with where each runs your apps and who it fits in 2026.

Eight Rancher alternatives for 2026 compared side by side, from OpenShift and Spectro Cloud to Portainer and Omni, with when each one fits.

Ankra's own platform now runs in Helsinki. The timeline, the bill, the benchmark numbers, and the places UpCloud let us down, from the team that did the move.

Moving a Next.js app on Vercel with a Supabase backend onto a Kubernetes cluster you own. What maps across, the database move, and the hard parts.

We ran the same k6, iperf3 and fio suite on EKS, GKE and Hetzner and lined the results up against UpCloud, OVH and DigitalOcean. The cheapest node in the test won more than it had any right to.

Two identical cloud VMs measured 2.8x apart on CPU; one failed etcd's fsync bar. A 15-minute in-cluster benchmark suite catches this before production does.

UpCloud vs DigitalOcean vs OVH on identical Kubernetes clusters. UpCloud led on storage and network, OVH on latency, DigitalOcean had the slowest tails.

ingress-nginx is retired and its successor was cancelled before it shipped. A migration path to Gateway API that does not require a big-bang weekend.

Half of all clusters are running an ingress controller that stopped receiving security patches in March. Meanwhile, teams debate service meshes. The five findings that show up in every real cluster security review, each with the incident that proves it and the fix.

Manage GitOps secrets by encrypting each value with SOPS and AGE before it reaches Git. Only ciphertext is committed and decrypted at deploy time.

New Ankra clusters no longer run ArgoCD. Here is the engineering story of the native deployment engine: why we built it, how the reconcile loop works, what the migration does to your live resources (nothing), and what we have not finished yet.

'Don't use Kubernetes until you have a platform team' was good advice in 2019. The premises changed. What a two-to-six person company should set up in week one, what to deliberately skip, and what it actually costs: about zero.

AI agents have wiped real production databases this year. Replit's went down during a code freeze, Cursor's took nine seconds via Terraform. Here's how Ankra's draft-first, GitOps-backed AI gives teams a senior DevOps agent without that failure mode.

Master the essential strategies and patterns for running reliable, scalable applications on Kubernetes in production. Learn from real-world examples and avoid common pitfalls.