DevOps, Kubernetes & cloud infrastructure
From completed code to a controlled release: we automate deployments, make systems observable and support your Kubernetes and cloud infrastructure.
15 minutes with Robert · free, no obligation
devs group builds and operates Kubernetes platforms for companies in Switzerland: in Google Cloud, at Swiss providers, or on-premise. We take over existing clusters as readily as we build on a green field, including CI/CD, observability and upgrades — as in the ERP rollout to 100 opticians for GMG.
DevOps for AI-assisted development
When coding agents produce more changes, tests, reviews and deployments need to keep up. We connect development with CI/CD, automated checks, traceable approvals and a planned rollback. Our AI-assisted software development service shows how this fits together.
Infrastructure that fits your system
Kubernetes and cloud infrastructure: we use the capable services of Google Cloud Platform (GCP) to build scalable, robust applications. For on-premise solutions we build on Kubernetes and bare-metal deployments, which give full control and maximum performance.
CI/CD: we automate your software delivery so new code is tested continuously and reaches production quickly.
Observability: we implement comprehensive monitoring, logging and tracing, so the performance of your systems is visible at all times and potential problems surface early.
Infrastructure as code: we define your entire infrastructure in code. That makes configurations repeatable and transparent, and produces a version-controlled, error-free environment.
Frequently asked
What does a managed Kubernetes cluster cost in Switzerland?
A production cluster with three worker nodes costs between CHF 158 and CHF 779 per month at Swiss providers; operating it comes on top and is usually the larger item. The full cost model, with price lists from Exoscale, Infomaniak and Flow Swiss, is in our article What Kubernetes really costs in Switzerland.
Do we need Kubernetes at all?
Not always. One application, one database and predictable traffic often run better and more cheaply on a well-sized VM. How to tell the difference is in Kubernetes is not a goal. We will tell you openly in a first conversation when a cluster is the wrong answer.
Do you take over existing clusters?
Yes. We audit the current state (versions, ingress, certificates, backups, alerting), plan the upgrade path, and then take on operations and support with defined response times. What a version jump without downtime looks like is in Kubernetes v1.37: upgrading without downtime.
Does our data stay in Switzerland?
If that is your requirement, yes: with Swiss cloud providers or on-premise deployments the entire stack stays in Switzerland. Where sovereignty technically breaks and where it does not, we worked through on the hardest case in Sovereign AI inference in Switzerland.
