See every node in a K3s cluster
Open 节点 / 机器 (Nodes / Machines) under 集群 (Cluster) in the Infra sidebar. After you choose an environment, it lists every node of that environment’s cluster with its Ready status, roles, internal and external IPs, CPU and memory capacity, operating system and Kubernetes version. Click a node to open its details.
The list covers the whole cluster. Application and namespace selections do not filter it.
What each column means
| Column | Source |
|---|---|
| Node | Node name |
| Status | The node’s Ready condition; cordoned nodes are marked as unschedulable |
| Role | Role labels such as control-plane or worker |
| IP | InternalIP and, when present, ExternalIP addresses |
| CPU / memory | Node capacity |
| OS / Kubernetes | OS image, kubelet version and architecture |
Use the search box to filter by node name or IP, and 刷新 (Refresh) to read the list again.
Several machines, one cluster
When several machines belong to one K3s cluster, a single cluster connection shows all of them. Independent clusters need one connection each, mapped to their own environment in project.yaml.
Permissions
The identity in your kubeconfig needs list on nodes for the list and get for node details. Nodes are cluster-scoped, so a namespace Role is not enough — grant a ClusterRole. See RBAC permissions.
Limits and errors
- Read failures are shown on the page; an empty list is never shown in place of an error.
- If the response hits the read limit, the page says the list is not the cluster’s full node set.
- For live CPU and memory usage per Pod, see Monitoring. Node usage metrics are not shown.
Frequently asked questions
Why does the Nodes page ignore my application filter?
Nodes belong to the whole cluster, not to an application or namespace. The page always lists every node returned by the cluster for the selected environment.