Managed Kubernetes, hosted in India
NABHCLOUD runs Kubernetes clusters on Apache CloudStack inside our Noida data centre. We set up the control node and the worker machines for you; you simply grab the kubeconfig file and start deploying. Clusters can run on an hourly meter with no lock-in, or on a monthly plan. All listed rates exclude 18% GST. Current plans and prices are shown on this page.
What's included with every cluster
No extra charge for the control node
We price a cluster purely by its nodes. Both the control node and each worker are charged at the same compute rate as our regular cloud servers, and the control node's cost is already folded into the plan price.
Working with kubectl right away
Grab the kubeconfig file from either the dashboard or the API, then use it with kubectl, Helm, Lens, k9s, or hook it into your CI pipeline exactly as you would elsewhere.
Scale as traffic shifts
Adjust vCPU and RAM on your nodes, or add and drop worker machines, straight from the dashboard or the API. No support ticket, no rebuilding a new cluster.
Pay hourly or lock in up to 36 months
Begin on the hourly meter with zero commitment, then shift to a monthly or longer-term plan once your usage settles, which brings the rate down. The billing cycle can be changed later at any point.
Hosted entirely in India
Your nodes and their NVMe storage sit in our Noida (Delhi NCR) facility, running on our own network and shielded by Cloudflare Magic Transit against DDoS traffic.
Billed in rupees, with a proper GST invoice
Fund your account through a prepaid wallet using UPI, a card, or net banking. Each invoice qualifies as a GST tax invoice, so you can claim input credit against it.
Getting from sign-up to kubectl
Register for an account, finish the one-time eKYC check, then open the Kubernetes section in the dashboard to spin up a cluster. Once its status reads Running, pull down the kubeconfig.
Step 1: Spin it up
Choose a cluster plan and a billing cycle, give the cluster a name, and create it. It's ready to use once the status shows Running.
Step 2: Hook up kubectl
export KUBECONFIG=$HOME/Downloads/kubeconfig kubectl get nodes kubectl create deployment web --image=nginx
Step 3: Expand later
Bring in more workers or bigger nodes from the cluster screen or the API whenever traffic rises, and switch from hourly billing to a fixed term once your load stabilises.
Only need a single machine for something like Docker Compose or a smaller app? Try our cloud VPS instead, or configure one to the exact specs you need.
Frequently asked questions about managed Kubernetes
What is managed here, and what do I manage?
We handle cluster creation on Apache CloudStack Kubernetes Service (CKS) — the control node, the worker nodes, their networking, and the Kubernetes install itself. You get the kubeconfig and run your own applications with kubectl, Helm, or whichever CI tool you prefer. Everything you deploy, including your container images and data, remains under your control to manage and upgrade.
Which Kubernetes version do you run?
We run Kubernetes through Apache CloudStack Kubernetes Service (CKS), hosted from our Noida (Delhi NCR) facility.
Is there a separate control plane or cluster fee?
No, there isn't one. A cluster's price is built from its nodes alone — the control node and each worker are billed at the same compute rate as our standard cloud servers, and the displayed plan price already covers the control node.
How does cluster billing work?
You can choose hourly metering with no commitment, or a monthly plan and longer terms running up to 36 months at a reduced effective rate. Switching a cluster between hourly and a fixed term is possible later on. All prices exclude 18% GST, and you'll receive a proper GST tax invoice for every charge.
Can a cluster be resized?
Yes. You can change node vCPU and RAM, or add and remove workers, through the dashboard or the API. On a fixed-term plan, the remaining paid period gets settled on a pro-rata basis; an hourly cluster simply starts billing at the new size going forward. Adding or removing workers happens on the live cluster without downtime, but changing a node's vCPU or RAM pauses the cluster for a few minutes while it rebuilds.
How do I connect using kubectl?
Pull the kubeconfig file from the cluster's dashboard page, or fetch it via the API, then point kubectl to it. From there it behaves the same with Helm, Lens, k9s, or any CI pipeline.
Do you offer GPU nodes or dedicated hosts for Kubernetes?
No. Every node runs as a KVM virtual machine with reserved vCPU and RAM backed by NVMe storage. We don't rent out GPU servers or physical dedicated hardware for Kubernetes.
Does my cluster stay located in India?
Yes. It runs from our facility in Noida, Uttar Pradesh, on our own Indian network and behind Cloudflare Magic Transit DDoS protection. Both your nodes and their storage remain within India.