Kubernetes
Create a managed Kubernetes cluster, connect with kubectl, and manage worker node pools.
Simplifyd Cloud manages the Kubernetes control plane. You choose the Kubernetes version and the worker nodes that run your workloads. Managed Kubernetes is in beta.
Create a cluster
- Open the project canvas and select the environment where you want the cluster. The workspace wallet must be funded.
- Open the command palette with
Cmd + KorCtrl + Kand choose Kubernetes Cluster. - Enter a Name and choose a Version. You can create a cluster on Kubernetes 1.34 or 1.33. Simplifyd Cloud manages the patch version.
- Configure at least one Node pool, then click Create.
- Open the new cluster on the canvas and click Deploy cluster on the Deployments tab. Provisioning takes a few minutes. When the deployment shows Running, click View nodes to check node and pod readiness.
Each cluster supports one to eight pools. Each pool contains identically sized workers. Give each pool a unique name of up to 32 lowercase letters, numbers, and hyphens. Start and end the name with a letter or number.
| Setting | Range |
|---|---|
| Nodes per pool | 1–32 |
| vCPUs per node | 1–32 |
| Memory per node | 2–128 GB |
| Root disk per node | 25–500 GB |
Worker vCPU, memory, and root disk are billed by the second. The control plane is included. The cluster's Settings → Billing section shows the total resources across its worker pools.
Connect with kubectl
Install kubectl for your cluster's Kubernetes minor version and install the Edge CLI. Make sure edge is on your PATH and sign in with edge auth login using an account in the cluster's workspace.
Open the cluster on the canvas and go to Settings → Access → Download kubeconfig. In the directory containing the file, run the following commands. Replace my-cluster with your cluster's name.
export KUBECONFIG=./my-cluster-kubeconfig.yaml
kubectl get podsYou can also generate a kubeconfig from the terminal. Run edge link and select the cluster's workspace, project, and environment. Copy the Service ID from the cluster's Settings and replace CLUSTER_SERVICE_ID in the command below.
edge cluster kubeconfig CLUSTER_SERVICE_ID --out ./my-cluster-kubeconfig.yamlBoth methods configure kubectl to call edge for a short-lived cluster token. Your workspace role determines your default Kubernetes permissions. Owners get the cluster-admin role, developers get edit, and billing members get view.
Deploy a workload
With the cluster's kubeconfig set, run this example as a workspace owner or developer. It creates an NGINX Deployment and a Kubernetes Service on port 80.
kubectl create deployment web --image=nginx:alpine
kubectl expose deployment web --port=80
kubectl rollout status deployment/web
kubectl port-forward service/web 8080:80Open http://localhost:8080 while port forwarding is running. Press Ctrl + C to stop forwarding.
Expose the workload over HTTPS
Open the canvas's Network view and find the cluster's private hostname. For a hostname such as k8s.p28xl.0000001.simplifyd.internal, choose a public host such as web.k8s.p28xl.0000001.apps.simplifyd.app. Use a single label for the app name (web in this example).
Create a Kubernetes Ingress that points that host to the Service. Replace the example hostname with your own.
kubectl create ingress web \
--rule='web.k8s.p28xl.0000001.apps.simplifyd.app/*=web:80'Simplifyd Cloud routes the hostname to the cluster and provisions its HTTPS certificate. If you specify an Ingress class in a manifest, use cilium.
To use your own domain, register it against the cluster. Use the cluster's Service ID and run the command from a directory linked to the cluster's environment.
edge service ingress add CLUSTER_SERVICE_ID --protocol http --port 80 \
--custom-fqdn web.example.comFollow the DNS instructions printed by the CLI, then use your domain as the host in the Kubernetes Ingress rule.
Change node pools or upgrade Kubernetes
In Settings → Node Pools, add or remove pools or change the node count, vCPUs, memory, and disk. Click Save changes to start the rollout. Removing a pool removes all its workers. A workload pinned to that pool needs another eligible pool before it can be rescheduled. To select a pool, use the node label node.cluster.x-k8s.io/node-pool with the pool's name as the value.
In Settings → Version, select Upgrade to when a newer minor version is available. Upgrades move forward one minor version at a time and replace nodes one at a time. Allow several minutes per node. Downgrades are unavailable.
Delete a cluster
Open Settings → Delete Cluster, click Delete cluster, and confirm. This permanently deletes the cluster, its nodes, and the workloads running on them. Back up any data you need before deleting the cluster.