Expert hosting for expert software engineers

VSHN, The DevOps Company from Zurich, operates APPUiO. VSHN and Puzzle ITC founded APPUiO together in 2016, and in July 2024 VSHN acquired the brand rights from Puzzle. Since then APPUiO is VSHN's container hosting for teams that build their own software, and VSHN's dedicated clusters are offered as Managed OpenShift.

Memory is measured per pod as a one-minute mean, starting at 128 MiB for one minute and then in 1 MiB steps, so the smallest pod costs about CHF 6 for a month. If a pod uses less than it requests, the request is billed. CPU is included up to the zone's memory-to-CPU ratio: in a zone with 4096 MiB per core, a pod requesting 1024 MiB and 0.4 cores is billed for 1638 MiB, because 0.4 cores correspond to 1638 MiB. The APPUiO portal lists the ratio for each zone.

A team running production plus a few test and preview environments lands around 20 GiB of memory, about CHF 980 a month, plus storage and any managed databases. The number depends on your requests and how long your preview environments run, so tell us what you run and we send an estimate based on your actual workloads.

APPUiO runs on the Baseline service level and comes without guarantees of resource availability. Community support is included, and paid support plans are available on top. If you need a contractual availability target, around-the-clock response or a cluster that is yours alone, that is Managed OpenShift, which VSHN operates under an SLA. Managed Kubernetes is still in development.

In two Swiss zones: Cloudscale in Lupfig (AG) and Exoscale in Geneva (GE). Each zone is one highly available OpenShift cluster in one data center of that provider, and VSHN operates both. You choose the zone for each project, and you can use both on the same account.

The data stays in the zone you choose, at Cloudscale in Lupfig or at Exoscale in Geneva. Who can access the systems is a separate question: VSHN operates APPUiO from Switzerland and Canada, and VSHN Canada carries out scheduled night-time work. If your data must never be reachable from abroad, run it on Managed OpenShift with the Switzerland-only support option, which rules out access from outside Switzerland by contract.

Yes, by deploying your application to both zones. APPUiO does not replicate between zones for you: each zone is a separate cluster, and managed databases replicate only inside one cluster. So the application runs in both places, and keeping data in step is part of your architecture. Exoscale object storage can replicate across zones, Cloudscale object storage cannot. For an engineered setup across sites, talk to us about Managed OpenShift.

By default the scheduler spreads the pods of a Deployment or StatefulSet across the servers of the cluster, so two or more replicas keep serving while a node is maintained. That default is a preference, not a hard rule. If two replicas must never share a server, set your own topology spread constraints or pod anti-affinity and define a PodDisruptionBudget. A single replica has no redundancy.

The K8up backup operator is preinstalled in every zone. You define which volumes are backed up, how often and to which S3 bucket, and K8up runs the backups and the restores. Managed PostgreSQL from the VSHN Application Catalog includes its own backups with point-in-time recovery, so database backups do not need a K8up schedule of their own.

Yes. HTTP and HTTPS go through the shared router at no extra charge. For TCP or UDP, such as a database port, MQTT or a game server, create a Service of type LoadBalancer. It gets a dedicated IPv4 address for CHF 20 a month and is available in the Cloudscale zone. Only IPv4 is supported for these addresses.

Privileged containers, and installing your own CustomResourceDefinitions. Your own operators can run, as long as they work on the objects that already exist in the cluster. If your stack depends on privileged containers or its own CRDs, a dedicated cluster with Managed OpenShift or Managed Kubernetes is the right fit, and we can help you size it.

There is no lock-in. APPUiO uses the standard Kubernetes APIs, so your images, manifests and Helm charts run on any other Kubernetes, and there is no minimum term: you can cancel at any time and are not charged for the next billing cycle. Backups made with K8up are plain restic repositories in your own S3 bucket.