Outgrow one server. Keep the same panel.
When one VPS is no longer enough, move the same apps to your provider's managed Kubernetes — DigitalOcean, Vultr or Linode. Replicas, autoscaling and rolling releases, deployed with the same git push. No YAML to write, no cluster to babysit.
Start on a $5 VPS. Move up when the traffic does.
One server
Push from Git and your app runs on a single VPS, with its database beside it. Cheap, simple, and enough for most products for a long time.
A few servers
Split the API, the workers and the database across machines in the same region, all from one panel, still on plain VPSes.
A cluster
Point Rubuz at a managed Kubernetes cluster and deploy the same apps there, with replicas across nodes and autoscaling on load.
Kubernetes power. None of the ceremony.
You keep the account with your provider and pay them for the nodes. Rubuz deploys and runs your apps on the cluster.
Replicas across nodes
Run several copies of each service on different nodes. Lose a node and the app keeps answering while the cluster replaces it.
Autoscaling
Add replicas when CPU or memory climbs and remove them when it falls, within limits you set.
Rolling releases
Each git push rolls out gradually and stops if new replicas fail their health checks, so a bad build never takes the site down.
Domains and TLS
Point a domain at the cluster and certificates are issued and renewed, the same as on a single server.
Secrets and environment
The variables you set in the panel are delivered to every replica as Kubernetes secrets.
One panel for everything
VPS apps and cluster apps side by side, with the same logs, alerts, team roles and billing.
The managed clusters developers already like.
DigitalOcean
DOKSDigitalOcean Kubernetes, with the control plane managed by DigitalOcean and nodes billed as Droplets.
Vultr
VKEVultr Kubernetes Engine across Vultr's regions, on the same compute you use for VPSes.
Linode (Akamai)
LKELinode Kubernetes Engine, with node pools you size and Akamai's network in front.
One region per cluster. Apps on a VPS — at these providers, at Hetzner, AWS, Google Cloud or anywhere else — keep working exactly as they do today.
or $249 / year. Unlimited clusters on that provider.
- Every Rubuz feature
- Replicas, autoscaling and rolling releases
- Domains, TLS and secrets on the cluster
- Your VPS apps in the same panel
Cluster nodes are billed by your provider.
Questions
Do I need to know Kubernetes?
No. You deploy the way you do on a VPS — git push or a Docker image — and Rubuz writes and applies what the cluster needs.
Does Rubuz create the cluster?
You create the cluster in your provider account, or let Rubuz create it with an API token you give it. Either way the cluster and its nodes stay in your account.
Can I mix VPS and cluster apps?
Yes. A database can stay on a VPS while the API runs on the cluster, or the other way round, all in one panel.
What about AWS and Google Cloud?
Apps on EC2 or Compute Engine VPSes work today, like any other server. Managed clusters start with DigitalOcean, Vultr and Linode.