Inside Private Cloud
Managed Kubernetes in the dedicated package: cluster design, migration, and 24/7 NOC with an architect.
Best for: Production clusters you need running now
Clear options. Clear when to use each
Self-serve clusters
Console SKU when the catalog unlocks: create a cluster, size the pools, pay per node.
Best for: Teams that want to start without a sales call
Who it's for
Platform teams
Microservices and internal platforms without a large SRE bench behind them.
Teams leaving DIY K8s
Keep the Kubernetes API you already build on, drop the weekend on-call for the control plane.
AI and batch on containers
Scale inference workers and batch jobs next to Optidata GPU capacity.
Migrations from hyperscalers
Same manifests, predictable bill, an architect on the migration.
What you get
Managed control plane
Optidata runs the API server, etcd, and scheduler. You get a kubeconfig, not a pager rotation.
Node pools that scale
Group workers by shape, grow a pool when demand climbs, shrink it when the peak passes.
High availability option
Redundant control plane for clusters that cannot afford a single point of failure.
Upgrades on your window
Versions and patches land on a maintenance window you choose, not a surprise Friday.
Native to the platform
Load Balancers front services, Block Storage backs volumes, VPC keeps east-west traffic private.
GPU path
Container workloads that need silicon land on the same RTX PRO 6000 Blackwell fleet as GPU Cloud.
No egress tax on scale
Traffic between your services does not arrive as a line item at the end of the month.
Human support behind it
24/7 NOC and engineers who answer, not a ticket queue that resolves itself.
Kubernetes vs Cloud Compute
| Kubernetes | Cloud Compute | |
|---|---|---|
| Unit | A cluster and its node pools | An instance |
| You manage | Manifests and workloads | OS, runtime, and the process supervisor |
| Scaling | Pool autoscaling and pod replicas | Resize or add instances |
| Best for | Containerized services that scale | Single apps, databases, legacy stacks |
From config to live
Define the cluster
Region, Kubernetes version, and network. Optidata provisions the control plane.
Size the node pools
Balanced or Dedicated shapes per pool, with min and max for autoscaling.
Connect and deploy
Pull the kubeconfig, point kubectl or your CI at it, apply your manifests.
Operate with the platform
Services get Load Balancers, PVCs get Block Storage, cost and state stay in one console.
Kubernetes pricing
Cluster management is a flat monthly fee for the control plane. Worker nodes are Cloud Compute instances with the same plans, storage, and add-ons, and that cost stacks on top of management. High availability is an optional add-on to the control plane.
| Item | Price |
|---|---|
| Cluster management | $20.00/mo |
| High-availability control plane | $40.00/mo |
The management fee alone is not the cluster bill. Worker nodes are Cloud Compute instances billed at Compute prices. Inside Private Cloud, Kubernetes is part of the dedicated package instead.
Prices from the Optidata console catalog · Miami, US (US1) · USD · updated Jul 30, 2026.
Short answers
Can I run Kubernetes on Optidata today?
Yes, inside Private Cloud. An architect designs the cluster and Optidata operates it. The self-serve console SKU is on the roadmap.
Who manages the control plane?
Optidata. You get cluster access and own your workloads.
Does it work with my existing manifests?
Standard Kubernetes manifests and Helm charts apply. Swap the ingress and storage classes for the Optidata equivalents.
Can nodes have GPUs?
GPU node pools are designed against the same RTX PRO 6000 Blackwell fleet as GPU Cloud. Availability locks with the SKU.
What about traffic between services?
Internal traffic stays on your VPC, and Optidata does not charge egress fees on it.