View, manage, and troubleshoot worker nodes from the page in the ACK console.
Navigate to the Nodes page:
Log on to the ACK console. In the left navigation pane, click Clusters.
On the Clusters page, click the name of your cluster. In the left navigation pane, click .
Connect to a node
For troubleshooting, performance monitoring, or running custom scripts, connect to the node's ECS instance.
In the Actions column, choose 
> Workbench Connection or > VNC Connection.
See Methods for connecting to an ECS instance for other options.
ContainerOS does not support direct logon or SSH access to prevent untraceable operations and reduce security risks. For maintenance, see work with the administrative container of ContainerOS.
Drain a node
Draining a node evicts all pods from the node and marks it as unschedulable, preventing new pods from being scheduled to it.
In the Actions column, choose > Drain and follow the prompts.
Before draining a node:
Ensure that other nodes have sufficient resources to accommodate the evicted pods.
Verify that pod affinity rules and scheduling policies allow rescheduling to other nodes.
DaemonSet-managed pods are not evicted during the drain operation.
Set node scheduling status
Use this operation to manually mark a node as unschedulable or schedulable, for example, before maintenance. Perform this operation during off-peak hours to minimize workload impact.
Select the node, then click Set Node Schedulability at the bottom. Read the precautions in the dialog box and follow the prompts.
Unschedulable nodes are labeled SchedulingDisabled. Existing pods continue serving traffic, but no new pods are scheduled. DaemonSet-managed pods are not removed.
Remove a node
Removing a node disconnects it from the cluster. You can choose whether to also release the associated ECS instance. Perform this operation during off-peak hours to minimize workload impact.
To remove a single node: in the Actions column, choose
> Remove Node and follow the prompts.To remove multiple nodes: select the nodes, click Batch Remove at the bottom, and follow the prompts.
See Remove a node for precautions.
Monitor node resources
Click Monitor in the Actions column. If the monitoring component is not yet installed, you are prompted to install it. After installation, Managed Service for Prometheus is enabled and a resource dashboard is available for the node.
Create custom PromQL alert rules for abnormal node conditions. See Best practices for configuring alert rules in Prometheus.
Diagnose node faults
Click Exception Diagnosis in the Actions column to trigger an automated inspection and receive a repair plan.
See Node diagnostics for supported scenarios and repair plans.
Manage node labels and taints
On the Nodes page, click Manage Labels and Taints to configure labels and taints that control pod scheduling.
Run batch operations on nodes
Apply the same action to multiple worker nodes at once, such as updating the OS kernel or installing monitoring, security, and audit packages.
Select the target nodes, click Batch Operations at the bottom, and follow the console guide.
Note: Batch operations are not supported on clusters with Auto Mode enabled.
View node details
View resource usage
In the Actions column, click Details to view node resource usage and status at a glance.
CPU and memory metrics:
Metric | Formula |
CPU request | Sum of CPU requested by all pods on the node / allocatable CPU of the node |
CPU utilization | Sum of CPU used by all pods on the node / allocatable CPU of the node |
Memory request | Sum of memory requested by all pods on the node / allocatable memory of the node |
Memory utilization | Sum of memory used by all pods on the node / allocatable memory of the node |
Allocatable resources = Resource capacity - Reserved resources - Eviction threshold. See Resource reservation policy.
The Details view also includes:
Basic information: node name, IP address, instance ID, container runtime version, operating system, and kernel version.
Resource allocation: CPU and memory requests and limits for all pods on the node
Node status: current condition and event history
Node events: recent event records
Pod list: all pods running on the node
View node configuration in YAML
In the Actions column, click Edit YAML to see the full node YAML, useful for deep inspection and troubleshooting.
Next steps
Use resource profiling to get container resource recommendations based on historical usage, then configure resource requests and limits when creating a Deployment.
Configure node labels and a node selector to schedule pods to specific nodes.
Upgrade or downgrade the configurations of a worker node to scale resources.
Attach data disks to nodes for additional storage for the container runtime and kubelet.
Node upgrades (kubelet and container runtime versions) are managed at the node pool level. Update a node pool to apply version upgrades.