Using the console
Viewing cluster resources
Go to CloudWatch Container Insights page and select the petsite (EKS cluster name) in the Filter by cluster drop down. Your screen should look like the one below. You should also be able to filter any specific object in your cluster just by using the Find resources textbox.
For example, if you only want to see the EKS Pods, just type pods and see the table getting filtered.
Map view
Click on Map view which will show the cluster resources in a tree format for easier understanding of resources in the cluster. To see some critical metrics about a resource, just click on the node to see the metrics showing up at the bottom of the screen. In the screenshot below, the cluster petsite has been selected and you can see CPU, Memory and Network related metrics aggregated at the cluster level.
Now click on View dashboard. You will be able to see the automatic dashboards created by CloudWatch Container Insights. For EKS, you can see performance and health metrics at the following levels:
- EKS Namespace
- EKS Cluster
- EKS Nodes
- EKS Service
- EKS Pods
Simply select the level at which you want to see the metrics as shown below and apply filters as required.
As you can notice in the above GIF, you can also see Application and Performance logs for a selected resource.
For example, simply select EKS Services, and select service-petsite from the Filters drop down.
In the Pod performance table under the dashboard, select one of the petsite-deployment entries and click on Actions and click Application logs or Performance logs which will take you to CloudWatch Logs Insights screen as shown below.
Notice that the log group automatically selected, timeline selected and a sample query also present there. Simply click on Run query to see the log data.