Skip to main content

Managed Swiss Cloud

We run our own cloud from Bern on Proxmox VE and Ceph, on our own hardware in a Swiss datacenter. It gives you managed VMs, managed Kubernetes and more than 40 managed services. You pay for what you use.

vCPUMemoryBlock StorageBlob StorageBackupPublic IPBandwidth
git push build image deploy
Fully managed by Natron
Deployment & Operations
cpu67%
mem72%
net47%
Backup completed: 2.1 TB
All health checks passing
Rolling update: api v2.4.1
Kubernetes Cluster
Your pods
web
2c · 4Gi
api
1c · 2Gi
worker
0.5c · 1Gi
+1 running
Managed services on Kubernetes
Harbor
ArgoCD
+6 more
ciliumingresscert-mgrprometheusgrafanaloki
Virtual Machines
Your VMs
app-1
4c · 8Gi
dev
2c · 4Gi
Managed services on VMs
PostgreSQL
Redis
+10 more
proxmoxha-clusterlive-migratebackupsnapshots
Proxmox VE + Ceph
Swiss Datacenter
firewall / vpn
fwvpnidssdw
switch-1
node-1
85%
node-2
62%
node-3
71%
node-4
78%
node-5
55%
node-6
68%
switch-2
25g fabric · nvme · 3x replicated · redundant power

Two ways into Natron Cloud

Both run on the same hardware and the same Kubernetes. The difference is who operates the platform.

Recommended

Managed by Natron

We run VMs, Kubernetes clusters in three tiers and managed services with an SLA of up to 99.9%. Your team deploys; we handle the platform, patching, backups and monitoring.

Book a call
Self-Service

Self-service with KlickOps

You connect a repository and get a URL. Domain, backups and a database are included, billed in Swiss francs from prepaid credit. KlickOps is currently in private beta.

Go to KlickOps

How we run Natron Cloud

Everything on this page is also in our code. Here are eight facts from operations.

Proxmox VE and Ceph

We run our own hardware with Ceph, replicated three ways with compression; storage and compute share the same cluster.

High availability with rules

Proxmox HA runs with affinity and anti-affinity rules. If a node fails, VMs restart on another.

Maintenance without downtime

Hosts are drained automatically, workloads live-migrated, and a node is updated only after a Ceph health check.

A network per customer

Every customer gets a dedicated VLAN zone in Proxmox SDN. The zone model is default deny, with dual-stack IPv6 and BGP.

Backups every night

VM backups go to dedicated Proxmox Backup Servers, Velero keeps Kubernetes backups for 30 days, and we monitor freshness.

Staged patching

We patch monthly: dev first, then staging, then production. Kernels and packages are versioned and pinned.

449 alert rules

Prometheus, Grafana and Loki send alerts in three priorities to on-call, backed by a dead-man's-switch.

SSO and least privilege

Every login to Proxmox and Kubernetes is personal, through single sign-on, with operator roles that carry only the rights they need.

Security and datacenter in detail

Customers after the migration

nicolas_wälti
Natron helped us migrate and consolidate our Kubernetes clusters to their centralized, managed and secure stack. They provided invaluable support throughout the migration process and beyond.
Nicolas Wälti

Nicolas Wälti

Head of SaaS Team, Inova Solutions

Frequently asked questions

Where is my data?

In a Swiss datacenter on hardware that Natron owns and operates. There is no foreign parent company and no hyperscaler in between. The operator is Natron Tech AG in Bern.

Who operates the hardware?

Our engineering team in Bern. We buy, rack and run the servers ourselves, including network, firewall and backup systems.

How is this different from KlickOps?

KlickOps is self-service: you connect a repository and get a URL, with database and backups included. Managed Natron Cloud means our engineers run VMs, Kubernetes clusters and services for you, with an SLA and a named contact. Both run on the same infrastructure.

How is it billed?

By consumption: vCPU and memory per hour, storage, backup and network per month. Managed Kubernetes and managed services are billed on top as operations services. You get a quote based on your workloads.

Which SLA applies?

Standard 99.5 percent, Enterprise with high availability 99.9 percent, both per month. Support tiers from best effort to a 4-hour response at 5x13; 7x24 on request. Details are on the service levels page.

How do I migrate from a hyperscaler or from VMware?

We assess your environment, build the target platform, migrate VM by VM or cluster by cluster and run both in parallel until you sign off. For VMware we use the Proxmox ESXi import.

Can I combine VMs and Kubernetes?

Yes. Databases run on VMs and applications on Kubernetes, in the same network with the same backup and monitoring. Many customers start with VMs and move workloads to containers step by step.

How do I leave?

The stack is open source: we export VMs as qcow2 or raw, object data over S3, databases as dumps, and your Kubernetes manifests live in your Git. There is no technical lock-in.

Free Guide|Infrastructure

Why Organizations Are Not Moving to Proxmox

  • 3 years of production Proxmox experience with 1000+ VMs
  • Honest assessment of what VMware migrations actually require
  • HA, monitoring, and Ceph storage lessons learned the hard way