Running Kubernetes on a network you can't configure means designing around someone else's VLANs, routing and firewall. On KASTERVO, you build your Kubernetes cluster on compute instances inside your own KVPC and set the VLANs, subnets, routing and edge firewall rules it needs. You install and run Kubernetes; we operate the hardware and the high-availability cluster underneath.
When instance types come in fixed shapes, nodes end up oversized in one resource to get enough of another. KASTERVO Compute Instances are priced per GB of RAM, per CPU core and per GB of flash storage, so each node gets the mix it needs.
Service IPs, node subnets and ingress traffic all depend on routing and firewall changes. When those settings sit behind a higher tier or a support ticket, every cluster change waits. In your KVPC, those settings are yours to change.
A Kubernetes API reachable from a shared network widens the attack surface. Put the control plane and nodes in their own VLAN, and expose only the services you choose through your edge firewall.
When the network under a Kubernetes cluster is out of your hands, engineers build around it with extra NAT layers, overlay workarounds and waits for provider changes. That time goes into plumbing instead of the platform you're building. In a KVPC, you configure the routing and firewall to fit the cluster design you want.
Run control-plane and worker nodes as separate compute instances in your KVPC. Our high-availability cluster sits under all of them; redundancy inside Kubernetes, such as three control-plane nodes, stays part of your cluster design.
Each node has a fixed monthly price set by its RAM, cores and storage. Adding a worker adds a known amount to the bill, however hard the cluster works.
Install the Kubernetes distribution and tooling you prefer on your nodes. The cluster, its upgrades and its workloads are yours to run.
Run containerd, Docker or another runtime on your compute instances, as you would on any Linux server you administer.
Deploy Prometheus or an ELK stack on compute instances alongside the cluster.
Run Jenkins or GitLab runners on their own compute instances, so pipelines deploy to the cluster over your private network.
The network underneath your compute instances is encrypted end to end, and customer data stays in the EU under GDPR.
Share your node count and network layout, and we'll size the KVPC and compute instances around them.
Get in touch
For MSPs and software houses that build and run Kubernetes platforms for their own clients.
See the program
Why we advertise Kubernetes LoadBalancer IPs over BGP with Cilium and pfSense instead of Layer 2 announcements.
Read the blog