Skip to main content
Latitude.sh Managed Kubernetes runs production-ready clusters on dedicated bare metal infrastructure. You get the performance benefits of bare metal combined with the convenience of a fully managed Kubernetes service.

​
Features

  • Managed control plane: Focus on your workloads while cluster infrastructure is handled for you.
  • Bare metal performance: Run workloads on dedicated hardware for maximum performance and predictable latency.
  • Full kubeconfig access: Connect to your clusters using standard Kubernetes tools and workflows.
  • Native LoadBalancer support: Built-in load balancing for your services via MetalLB and BGP.

​
Requesting access

Managed Kubernetes is currently in preview.
1

Navigate to Kubernetes

Log in to the dashboard, select a project, and navigate to Kubernetes in the sidebar.
2

Fill out the request form

Complete the early access request form with your expected cluster count and use case.
3

Submit your request

Submit the form and the team will review your request.

​
Creating a cluster

Create Kubernetes clusters from the dashboard. Clusters are deployed with Rancher RKE2.
Before creating a cluster, make sure you have a verified account with a payment method, at least one SSH key added to your team or project, and a project selected.
1

Access cluster creation

Log in to the dashboard, select a project, navigate to Kubernetes in the sidebar, and click Create cluster.
2

Select a location

Choose a data center region for your cluster. Available locations include Ashburn (ASH), Dallas (DAL), and Frankfurt (FRA). Plan availability and pricing may vary by location.
3

Configure control plane

Choose a control plane size based on your workload:
  • Small: Development and lightweight clusters
  • Medium: Staging and general workloads
  • Large: Production and larger workloads
Use the node count selector to configure 1-3 control plane nodes.
4

Add worker nodes (optional)

Click Add worker node to add compute capacity for your applications. Select a plan and configure the node count (up to 3). All worker nodes share the same plan.
5

Configure access and details

Select one or more SSH keys for node access. These keys let you SSH into the underlying bare metal servers.Enter a name for your cluster (3-63 lowercase characters, numbers, or hyphens).
6

Create the cluster

Review the pricing summary and click Create cluster.
After creation, you’ll be redirected to the cluster overview page. Provisioning typically takes several minutes.

​
Viewing your clusters

Navigate to Kubernetes in the sidebar to view your clusters. The cluster list displays:
ColumnDescription
NameThe cluster identifier
VersionThe Kubernetes version running on the cluster
Control PlaneNumber of control plane nodes
WorkersNumber of worker nodes
LocationThe data center region where the cluster is deployed
EndpointThe API server endpoint URL for connecting to the cluster
StatusCurrent cluster state (Provisioning, Ready, Deleting, or Error)

​
Managing your cluster

Click a cluster in the list to view its details and access configuration options.

​
Cluster access

From the overview page, you can download your kubeconfig file to connect to the cluster using kubectl:
  • Download kubeconfig: Save the file to your machine
  • Copy to clipboard: Copy the kubeconfig contents directly
  • View: Preview the kubeconfig in the dashboard
Connect to your cluster: 
kubectl --kubeconfig=./my-cluster-kubeconfig.yaml get nodes
The kubeconfig is available once the cluster reaches Ready status.

​
Nodes summary

View a summary of your control plane and worker nodes, including the total count and plan used for each node type.

​
LoadBalancer IPs

Managed Kubernetes clusters include LoadBalancer IPs announced via BGP using MetalLB. When you create a LoadBalancer service, an IP from this pool is assigned automatically.

​
Viewing cluster nodes

Navigate to the Nodes tab to view information about your cluster’s servers. The page displays two sections:
  • Control Plane Nodes: Servers running Kubernetes core services that maintain cluster state
  • Worker Nodes: Servers running your application workloads (pods and services)
Each section shows replica and ready counts in the header, with a table listing each node’s server name (linked to the server details page), IP address, and status. Click any row to open a details panel with additional information including the node type, internal IP, and external IP.

​
Deleting a cluster

To delete a cluster, go to the Settings tab and click Delete cluster in the Danger Zone section. You’ll need to type the cluster name to confirm.
This permanently destroys all nodes, releases all LoadBalancer IPs, and deletes all data and workloads.