Kubernetes Infrastructure: Nodes

Table of Contents

Nodes screen
Table columns
Node detail pane

A node is a worker machine in a Kubernetes cluster, running the container runtime, kubelet, and kube-proxy alongside your workloads. The Nodes screen lists every node across every monitored cluster, so you can spot capacity pressure or scheduling problems before they show up as pod failures.

Nodes screen

Click Infrastructure in the top navigation, then Nodes in the left sidebar.

Table columns

Each row is one node. Columns:

Column What it shows

Column	What it shows
Cluster	The cluster the node belongs to.
Name	The node’s name.
Status	The node’s kubelet health condition: `Ready` (healthy), `NotReady` (unreachable or unhealthy), or `Unknown` (state can’t be determined). This reflects node health only — a cordoned node can still show `Ready` while rejecting new pods; check Schedulable for that.
Schedulable	Whether the node currently accepts new pods — `NO` means the node is cordoned.
Age	Time since the node joined the cluster.
Kubelet	Version of the `kubelet` process running on the node.
Roles	Node roles, such as `control-plane` or `worker`.
CPU Cap. / CPU Allocable / CPU Used	Total CPU cores on the node, the capacity available for scheduling pods, and real-time usage.
Mem Cap. / Mem Allocable / Memory	Total memory on the node, the amount available to pods after system reservations, and real-time consumption by pods and system processes.

Cluster

The cluster the node belongs to.

Name

The node’s name.

Status

The node’s kubelet health condition: Ready (healthy), NotReady (unreachable or unhealthy), or Unknown (state can’t be determined). This reflects node health only — a cordoned node can still show Ready while rejecting new pods; check Schedulable for that.

Schedulable

Whether the node currently accepts new pods — NO means the node is cordoned.

Age

Time since the node joined the cluster.

Kubelet

Version of the kubelet process running on the node.

Roles

Node roles, such as control-plane or worker.

CPU Cap. / CPU Allocable / CPU Used

Total CPU cores on the node, the capacity available for scheduling pods, and real-time usage.

Mem Cap. / Mem Allocable / Memory

Total memory on the node, the amount available to pods after system reservations, and real-time consumption by pods and system processes.

Use COLUMNS to add or remove columns, and Download to export the current table.

Node detail pane

Click any row to open the node’s detail pane. It follows the same shared detail-pane layout as every other resource kind, with five tabs:

Details: Cluster, Name, Status, Schedulable, Age, Kubelet, Roles, and the CPU/Memory capacity figures from the table above, followed by the node’s Labels and Tags.
YAML: The node’s live manifest, as Structured (collapsible tree) or Raw text.
Metrics: Node-level CPU, memory, network, and disk usage as time-series charts. Use the chart time selector to zoom between recent activity (last 5m, 1h) and longer trends (24h).
Logs: System and application logs from pods running on this node, with full-text search, tag selection, and time-range filters.
Events: Kubernetes events for the node — readiness transitions, pressure conditions, and scheduling changes — each with a timestamp and source controller. Combine with Logs and Metrics to diagnose a NotReady node or a scheduling failure.