Partner offer · Managed Databases
Your customer doesn't buy “Kubernetes”. They buy managed PostgreSQL from you, powered by Syself + Hetzner, with replicated databases on local NVMe and none of the hand-built servers.

How it works
Immutable Kubernetes nodes and fast local storage usually pull in opposite directions. On Syself they don't, which makes it a fit for PostgreSQL, ClickHouse and other replicated stateful systems.

Syself exposes the Bare Metal disks with TopoLVM, so every database instance gets local-disk performance through standard persistent volumes.

With CloudNativePG, for example: replication, failover, backups, point-in-time recovery and the whole database lifecycle, declared in Kubernetes.

Data disks are separate from the OS lifecycle. Syself can reprovision the immutable OS for upgrades and repairs while the configured data disks stay untouched.
Immutable OS
v1.33 v1.34
Draining workloads Reprovisioning the OS Back in the cluster
Data disk
nvme0n1
Untouched
Data disk
nvme1n1
Untouched
Syself reprovisions the OS for upgrades and repairs. The configured data disks, and your customer's databases on them, stay as they are.
Who runs what
Your team owns the database service and the operator that runs it. Syself owns Kubernetes and the local storage layer, and Hetzner provides the disks and the servers.
PostgreSQL / ClickHouse
The service your customer buys
Database operator
e.g. CloudNativePG
Kubernetes
Provisioned, upgraded and healed
TopoLVM
Local disks as persistent volumes
Local NVMe
Separate from the OS lifecycle
Hetzner Bare Metal
Dedicated servers
What you sell
You set the SLAs, the price and the support model. Syself keeps the Kubernetes and machine lifecycle running underneath, and Hetzner provides the hardware at cost.






Our docs cover the best practices your team needs to run databases on Kubernetes.
Tell us about your company and the customers you serve. We reply within two business days to plan your first offer together.
