Pods are the smallest deployable units in Kubernetes. A pod runs an instance of independent application in Kubernetes. The instance may consist of one or more containers that are tightly coupled. This topic describes how to view details and modify configurations of a pod and manually scale container applications in the Container Service for Kubernetes (ACK) console.
Prerequisites
A workload is created. For more information, see Workloads.
View pods
View details of a pod
Log on to the ACK console. In the left-side navigation pane, click Clusters.
On the Clusters page, click the name of the cluster that you want to manage and choose in the left-side navigation pane.
On the Pods page, find the pod that you want to view and click View Details in the Actions column.
NoteOn the Pods page, you can modify and delete pods. For pods that are created by using a Deployment, we recommend that you use the Deployment to manage the pods.
The following table describes the status of pods.
Status
Description
Initialized
All init containers are started.
Ready
The pod is able to serve requests and added to the load balancing pools of all matching services.
ContainersReady
All containers in the pod are ready.
PodScheduled
The pod has been scheduled to a node.
For more information, see Pod Lifecycle.
View pod logs
On the Pods page, find the pod whose logs you want to view, and click Logs in the Actions column to view the logs.
Filter pods
On the Pods page, you can filter pods by name, label, pod IP address, and node IP address. You can also filter pods by keys and values that are displayed in the Label column.
Diagnose a pod
On the Pods page, find the pod that you want to diagnose and click Diagnose in the Actions column. For more information, see Work with cluster diagnostics.
Modify the upper and lower limits of CPU and memory resources for a pod
After you create an application, you can modify the upper and lower limits of CPU and memory resources for the application pods based on your business requirements. In this example, a Deployment is used.
Log on to the ACK console. In the left-side navigation pane, click Clusters.
On the Clusters page, click the name of the cluster that you want to manage and choose in the left-side navigation pane.
On the Deployments page, select a namespace from the Namespace drop-down list. Find the application that you want to manage and click Edit in the Actions column.
On the Edit page, configure the Resource Limit and Required Resources parameters to modify the upper and lower limits of CPU and memory resources. Then, click Update.
ImportantThe maximum number of pods that are supported by a cluster node is related to the network plug-in that is used by the cluster. For more information, see Compare Terway modes section of the "Work with Terway" topic.
Modify the configurations of a pod
Log on to the ACK console. In the left-side navigation pane, click Clusters.
On the Clusters page, click the name of the cluster that you want to manage and choose in the left-side navigation pane.
On the Pods page, find the pod that you want to modify and click Edit in the Actions column.
In the Edit YAML dialog box, modify the configurations and click Update.
Manually scale pods for an application
After an application is created, you can scale the pods that are provisioned for the application based on your business requirements.
Log on to the ACK console. In the left-side navigation pane, click Clusters.
On the Clusters page, click the name of the cluster that you want to manage and choose in the left-side navigation pane.
- Select the namespace where the Deployment is deployed, find the Deployment, and then click Scale in the Actions column.
In the Scale dialog box, set the Desired Number of Pods parameter to 4 and click OK.
NoteBy default, Deployments in Kubernetes are updated in rollingUpdate mode. This ensures that a minimum number of pods are available during the update. You can modify the minimum number of available containers in the template.
References
For more information about how to handle exceptions of pods, see Pod troubleshooting.
For more information about the best practices for pod scheduling, see Spread Elastic Container Instance-based pods across zones and configure affinities.