I build and run your clusters, migrate apps into k8s, and set up GitOps, autoscaling and monitoring. You get a reliable platform — without hiring a dedicated platform team.
The usual picture when a cluster gets spun up but no platform is built around it
Manifests scattered everywhere, deploys done from the terminal, nobody really knows what's running in prod. Recreating the cluster from scratch is impossible.
Nodes sit idle, requests/limits are guessed, managed control-plane and traffic add up. You pay several times what the cluster actually uses.
CrashLoopBackOff, OOMKilled, evicted — without proper metrics and logs, debugging becomes guesswork and incidents drag on for hours.
RBAC of "cluster-admin for everyone", no NetworkPolicy, unscanned images, secrets in ConfigMaps. One compromised pod and everything is exposed.
Under peak the service falls over, autoscaling is missing or misconfigured. And at idle the same resources burn budget for nothing.
The cluster exists, but there's no CI/CD for it, no ingress/cert-manager, no backups or upgrade plan. k8s is running but never became a platform.
The full lifecycle around Kubernetes — from your first cluster to a mature platform
I design and stand up a production-ready cluster — in the cloud or on your servers. Not bare k8s, but a ready platform with ingress, TLS and logs out of the box.
I move apps from VMs, docker-compose or another cloud into k8s — with zero downtime for users and without rewriting everything.
The cluster lives in git, deploys go through pull requests. ArgoCD/Flux keep state in sync, and any deploy rolls back with a single commit.
I make the cluster transparent: metrics, logs and traces in one place, clear dashboards and alerts before prod goes down.
The cluster grows under load and shrinks at idle. Requests/limits are set from real usage, not guesses — and the cloud bill drops.
I take the cluster under management: version upgrades, backups, incident response. Your platform engineer on a retainer.
Transparent packages with a fixed outcome. No abstract "billable hours"
A review of your existing cluster: reliability, security, cost and what to fix first
A production cluster with GitOps, ingress, monitoring and autoscaling — from zero to a working platform
Your external platform engineer: I keep the cluster healthy and respond to incidents
Targeted setup, cluster architecture reviews and Kubernetes consulting
From "we spun up a cluster" to a mature platform you can trust with prod
I review your current cluster or task, gauge load, reliability and cost, and find the bottlenecks
I fix the architecture and what to implement first. Transparent estimate and timeline
I stand up the cluster, write IaC, set up GitOps and monitoring. In sprints, with demos
I hand over docs, train the team and stay available. The cluster runs without surprises
A measurable outcome instead of "well, the cluster kind of works"
Cloud-native, open tools and CNCF standards — no vendor lock-in
Not always. If you run one or two services with no growth plans, k8s may be overkill — and I'll say so honestly at the audit. Kubernetes pays off when you need scaling, high availability, many services or teams. Sometimes k3s or even docker-compose is the right call.
Yes, that's a typical job. I start with an audit: find reliability, security and cost issues, get the cluster in order and build the missing pieces — GitOps, monitoring, backups. Without recreating from scratch unless it's actually needed.
I work with both. Managed (EKS, GKE, Yandex Managed Kubernetes) means less control-plane hassle. Self-hosted (kubeadm, k3s) is cheaper and provider-independent. At the audit we'll pick the option that fits your budget and requirements.
Cluster audit from €800, turnkey platform from €2,400, operations from €1,000/month. The exact price is fixed after a free express audit. You pay for results in stages — no hourly "meter".
Yes — a common, fast-paying-off request. Right-sizing requests/limits, autoscaling, spot nodes and removing idle resources typically cut the cluster bill by 30–50%. You get a report with concrete savings.
The opposite. Everything is described as code (IaC + GitOps) and documented, the stack is open and portable. I train your team so the cluster lives without me. Operations are optional, not a dependency.
Leave a contact — I'll run a free express audit and send what to improve in your cluster's reliability, security and cost
I'll reply shortly. No spam, no pushy calls.