Kubernetes
Cluster design, hardening and day-to-day operations on the managed clouds, or a complete platform on plain Linux servers with Kipper. We tell you which one fits, and we build both.
Managed cloud or your own servers
A managed cloud cluster and a Kipper installation on your own servers are both Kubernetes. They differ in cost, control and who carries the pager.
| Topic | Managed cloud cluster (EKS, GKE, AKS) | Kipper on your servers |
|---|---|---|
| Best for | Teams already deep in one cloud, spiky scale, strict compliance frameworks that name a provider | Teams that want predictable workloads, EU data residency and cost control on servers they own |
| Typical monthly cost | Control plane fee plus nodes, load balancers, storage and egress. A small production cluster commonly lands at $1,000 to $3,000 a month including compute (industry estimates, 2026) | The servers only, typically three mid-sized cloud nodes at the provider's list price, with SSL, storage and monitoring included by Kipper |
| Set-up time | Typically days, most of it IAM, networking and the add-ons listed below | One command for the platform, then the same care for your apps and data |
| What is included | The control plane. Ingress, certificates, storage, monitoring and backups are yours to assemble | Traefik ingress, Let's Encrypt, Longhorn storage, Prometheus, Grafana, Loki, backups, web console, SSO |
| Lock-in | Provider-specific add-ons accumulate quickly | Apache 2.0 software on standard Linux. Move the servers, keep the platform |
| Who runs it | You, or Labb on a retainer | You, or Labb on a retainer |
Three ways to start
Every one of them ends with documentation you own. Starting prices are the ones we invoice. Anything larger gets a written estimate before work starts.
One week in one cloud account or cluster. Security, cost, reliability and a prioritized to-do list with estimates. The report is yours to keep and to build with, whoever does the building. Larger estates are quoted after a short call.
A new platform, a move between clouds, or the move from a data center. Delivered in milestones with a named engineer, infrastructure as code and a hand-over your team can operate.
We keep the platform patched, backed up and observed, and respond to incidents within business hours at an agreed response time. Improvement work is billed by the day on top, so the monthly fee stays predictable.
Tools we reach for
Before you call
Possibly. If you run one or two services with steady load, a hardened server with Docker Compose and backups may be all you need. If you have several services, need zero-downtime deploys or want staging environments per branch, Kipper makes Kubernetes cheap enough to be the sensible choice.
Yes. The workloads are containers either way; what changes is ingress, storage, secrets and CI. We migrate one environment at a time, run both in parallel, and cut over when the metrics match.
Within the operations retainer, Labb responds during business hours at an agreed response time. Around-the-clock on-call is quoted separately and only where the workload justifies it.
Kubernetes releases three times a year and clusters fall behind quickly. Upgrades are part of the retainer, rehearsed on the test environment first.
Talk to an engineer
Thirty minutes with a senior engineer. Reply within one business day.