kloudedocs
Browse documentation
On this page

PRODUCT GUIDE

Clusters

Create and manage managed Kubernetes clusters in the Kloude portal: node pools, kubeconfig, Apply Manifest, Cloud Shell, and in-browser workloads, with a step-by-step guide for everyday operators.

A cluster on Kloude is a managed Kubernetes control plane plus the worker node pools you size yourself. You pick the region, Kubernetes version, node sizes, and a CNI (cilium or calico). Kloude runs the control plane; you run workloads with kubectl, GitOps, or the tools built into the portal.

Open Clusters from the portal navigation when you need capacity, a kubeconfig download, a quick look at Deployments and Pods, or a YAML apply without leaving the browser.

What you see in the panel

The list page shows each cluster's name, status (for example ControlPlaneStarting while it provisions), node count, version, region, network plugin, hourly cost, and created time. You can search by name, clear sorting, and refresh. The main action is + Create cluster.

Opening a cluster takes you to the detail view: Download kubeconfig, Apply Manifest, cards for node groups, status controls, the connected VPC, version, and CNI plus pod and service subnets. On the side you move between Overview, Node groups, and Cloud Shell (k9s), and you can search a list of Kubernetes resources, Deployments, Pods, Services, Ingress, PVCs, Config Maps, Secrets, and more.

When to use it

Use Clusters when you want managed Kubernetes for microservices, batch jobs, or stateful apps; when you need to grow or shrink node pools as demand changes; or when you prefer to inspect workloads from the portal instead of switching tools.

Tutorial: create and connect

Open Clusters

Select Clusters in the portal navigation and confirm the selected workspace. Prepare a VPC through Networking before starting cluster creation.

Create a cluster

Click + Create cluster. Choose a clear name, region (for example Milad), Kubernetes version, network plugin (cilium or calico), VPC, and at least one node pool size and count.

Wait until the control plane is ready

Back on the list, watch Status. While the control plane starts you may see ControlPlaneStarting and a node count like 0/3. Refresh until the cluster is ready before you rely on kubeconfig or workloads.

Open details and download kubeconfig

Click the cluster name. On Overview, use Download kubeconfig, then verify locally:

bash
kubectl --kubeconfig=./kubeconfig get nodes kubectl --kubeconfig=./kubeconfig get pods -A

Apply a manifest (optional)

Use Apply Manifest on the cluster page to paste or upload YAML when you do not want to switch tools. Prefer GitOps or CI for production changes.

Use the in-portal console

Browse Deployments, Pods, Services, Ingress, Events, and related resources from the side list. Open Cloud Shell (k9s) for interactive diagnostics. Prefer least-privilege kubeconfig for automation.

Scale or delete carefully

Manage Node groups when capacity changes. Delete a cluster only after workloads and dependent disks or IPs are no longer required, idle capacity still meters on Billing.

Overview tab

The Overview tab summarizes status, region, version, connected VPC, and high-level health so you can confirm the cluster is ready before you deploy.

Node pools

Node pools are groups of worker nodes that share a size. Scale a pool up or down, or add another pool when you need a different SKU.

Cloud Shell

Cloud Shell (k9s) opens an in-browser terminal against the cluster. Prefer kubeconfig with least privilege for automation; reach for Cloud Shell when you want interactive diagnostics.

Workload resources in the portal

The resource list lets you browse Deployments, Pods, Services, Ingress, Persistent Volume Claims, Storage Classes, Config Maps, Secrets, Service Accounts, Network Policies, and Events without kubectl. Pair PVCs with Disks. LoadBalancer Services tie into Networking. Combine Network Policies with Firewall rules on public IPs.