Skip to Content
Open Beta — managed Kubernetes, DNS & more are live.

Managed Kubernetes, end to end.

HA control planes, worker autoscaling, in-place upgrades, and integrated load balancers and volumes — hosted in Montréal, billed in CAD.

See it run

From an empty project to kubectl

The console as you will use it — pick a version and a size, watch the control plane and the workers come up, then connect with your own kubeconfig.

console.patrii.cloud 1:40

Creating a cluster, end to end — no cuts.

The whole cluster, wired for you

One command away from production

You pick the version and the sizes. We stand up the HA control plane, the worker pool, the load balancer, the volumes and the public entry point — and keep them healthy.

HA control plane ×3 / ×5 Autoscaling workers Octavia load balancer Persistent volumes Public or private
Simple pricing · Beta

Pay for the cluster, size the nodes

A flat control-plane fee, a small per-node management fee, and workers billed as ordinary VPS.

Cluster (control plane)

$9.99 /cluster·mo

HA control plane — run 1, 3 or 5 masters for high availability.

Node management

$4.99 /node·mo

Managed lifecycle per worker — upgrades, autoscaling and health.

Worker nodes

from $5.25 /mo

Any VPS flavor · you choose CPU / RAM

Workers are ordinary Patrii VPS — pick the flavor that fits.

Beta · all prices in CAD · no egress fees · deploy once your quota is approved.

From zero to kubectl in minutes

1

Pick a version & sizes

Choose the Kubernetes version, the control-plane size (1/3/5) and a worker flavor and count.

2

We provision the cluster

HA control plane, worker pool, load balancer, floating IP and storage classes are stood up for you.

3

Download your kubeconfig

Get an RBAC-scoped kubeconfig for you or your team — namespace or cluster wide.

4

Deploy, scale, upgrade

Apply your manifests, autoscale node groups, and upgrade in place when a new release lands.

v1.31 v1.32
1.31 1.32 node-1
1.31 1.32 node-2
1.31 1.32 node-3
rolling upgrade · workloads stay up
In-place upgrades

New Kubernetes versions, no maintenance window

Move the control plane and your workers to a newer Kubernetes release with a rolling, node-by-node upgrade — the pool is never fully down, and your workloads reschedule as each node comes back.

  • Auto — we roll out patch releases on a cadence you choose.
  • Frozen — you stay pinned and upgrade on your own schedule.
  • Rollback-aware — control-plane snapshots before every jump.
RollingZero-downtimeScheduled
Control plane

HA control planes that autoscale

Run 1, 3 or 5 control-plane nodes for high availability, and let the worker pool scale out automatically as demand grows — on dedicated nodes you size yourself.

1 / 3 / 5 mastersAutoscalingDedicated nodes
Create a cluster
Batteries included

Load balancers & volumes, wired up for you

Every cluster comes with load balancers and persistent volumes provisioned automatically — Service and PersistentVolumeClaim just work, no glue code.

LoadBalancer ServicePersistent volumesCSI & CCM
See the console
Day-two ops

In-place upgrades, RBAC kubeconfigs

Roll Kubernetes and node versions in place, and hand out RBAC-scoped kubeconfigs — then connect with kubectl, Helm and the tooling you already use.

In-place upgradesRBAC kubeconfigkubectl & Helm
Read the docs

Kubernetes questions

Yes. Clusters run conformant upstream Kubernetes provisioned with OpenStack Magnum on dedicated VPS nodes — no forks, no proprietary API. Your manifests, Helm charts, Operators and kubectl work exactly as they do anywhere else.

You choose 1, 3, or 5 control-plane nodes. With 3 or 5 masters, etcd runs as a quorum and the API server stays available through a node failure or a rolling upgrade. You can resize the control plane later without recreating the cluster.

Upgrades are in place and rolling: the control plane is upgraded first, then workers node by node, so the pool is never fully down and workloads reschedule as each node returns. You can let us apply patch releases automatically, or freeze on a version and upgrade on your own schedule.

Yes. Every cluster is wired to an Octavia load balancer with a public floating IP, so Services of type LoadBalancer and ingress controllers get a real external VIP. You can also run a fully private cluster with no public entry point.

Clusters ship with Ceph-backed storage classes, so PersistentVolumeClaims are provisioned dynamically and replicated three times in our Montréal region — no manual volume wiring.

You mint RBAC-scoped kubeconfigs per user or group, scoped to a namespace or the whole cluster, and revoke any of them in a click — without sharing a single god-mode admin kubeconfig.