Kubernetes Node Management

The Node Management page provides an overview and management interface for all nodes in a cluster, including list browsing, status filtering, scheduling control, and detail inspection.

Node List

The list page displays all nodes in a table with the following columns:

Column	Description
Name	Node name; click to open the detail page
Role	Control Plane or Worker, displayed as a badge
Status	Ready (green) or Not Ready (red)
IP Address	The node's IP address
K8s Version	Kubernetes version running on the node
CPU Capacity	CPU resource capacity of the node
Memory	Memory resource capacity of the node
Pressure	Memory, Disk, and PID pressure status; shown as red badges when abnormal
Age	Time since the node joined the cluster

Filtering and Search

The top of the list page provides filtering and search options:

• Role filter: All Roles / Control Plane / Worker
• Status filter: All Status / Ready / Not Ready
• Name search: Supports fuzzy search by node name and IP address

Nodes are cluster-scoped resources and do not belong to any namespace, so the namespace filter is not available on the node list page.

Node Actions

Each node row has a dropdown menu on the right side providing scheduling control actions:

Action	Description
Cordon	Mark the node as unschedulable, preventing new Pods from being scheduled to it. Already running Pods are not affected
Uncordon	Re-enable scheduling on the node, allowing new Pods to be scheduled

Cordon does not evict or stop running Pods. To safely evict all Pods from a node, use the kubectl drain command.

Node Detail Page

Click a node name to open the detail page. The header shows the node name, role, and Cordon/Uncordon action buttons. Content is organized into the following tabs:

Overview

The Overview tab presents node information in card format.

Node Information

Field	Description
Kubelet Version	The kubelet version on the node
OS Image	Node operating system image (e.g., Ubuntu 22.04.3 LTS)
Kernel Version	Node kernel version
Container Runtime	Container runtime type and version
Kube-Proxy Version	The kube-proxy version
Operating System	Operating system type (e.g., linux)

Creation Information

Field	Description
Created At	Time the node joined the cluster
Age	Duration since creation
UID	Unique identifier for the node

Capacity

Displays the node's total resource capacity (Capacity) as a card, including CPU, memory, Pod count, and other metrics.

Allocatable

Displays the node's available resources (Allocatable) as a card, including CPU, memory, Pod count, and other metrics.

Addresses

Lists the node's network addresses, including types such as InternalIP, ExternalIP, and Hostname.

Labels

Displays all current Labels as badges. Click "Edit" to enter edit mode:

• Add, modify, and delete key-value pairs
• Keys cannot be empty or duplicated
• Changes take effect immediately and may affect Pod scheduling

Conditions

Displays the node's status conditions. Each condition includes:

Field	Description
Type	Condition name (e.g., Ready, MemoryPressure, DiskPressure, PIDPressure)
Status	True (green), False (red), or Unknown (yellow)
Reason	Reason for the condition change
Message	Detailed description of the condition
Last Transition	Time of the last status change

The MemoryPressure, DiskPressure, and PIDPressure conditions correspond to the pressure indicators on the list page and can help diagnose resource bottlenecks.

Taints

Displays the node's taint configuration. Each taint has a Key, Value, and Effect field. Supported Effect values:

Effect	Description	Badge Color
NoSchedule	Pods without a matching toleration will not be scheduled to this node	Red
PreferNoSchedule	The scheduler tries to avoid scheduling Pods to this node, but does not guarantee it	Gray
NoExecute	Pods without a matching toleration will be evicted	Default

Taints can be edited: add or remove entries, each requiring a Key, optional Value, and Effect. Changes take effect immediately and may affect running Pods.

Events

Displays recent events related to the node, including timestamp, type (Normal / Warning), and message content.

For historical events, use command-line tools such as kubectl.

Related Documentation

• Cluster Overview
• Operations