Day 14 - Monitor performance of EKS Cluster using CloudWatch Container Insights

Welcome to Day 14 of 100 Days of AWS. The topic for today is Monitor performance of EKS Cluster using CloudWatch Container Insights.

Elastic Kubernetes Service(EKS) is the defacto standard for managing the Kubernetes cluster. Using EKS, AWS will take care of most tasks like patching and provisioning your cluster, performing upgrades, etc.

Using CloudWatch container insight, you can collect, summarize, and aggregate metrics and logs for your containerized application and microservices. It currently supports services like:

Amazon Elastic Container Service(ECS)
Amazon Elastic Kubernetes Service(EKS)
Kubernetes platforms on Amazon EC2

These metrics include CPU, Memory, Disk, and network utilization. It also helps us provide diagnostic information like container restart failure to help you quickly isolate and resolve problems.

To launch your Kubernetes Cluster, you can use tools like eksctl. For more info check the following link.

To deploy CloudWatch Container Insights run the following command

curl https://raw.githubusercontent.com/aws-samples/amazon-cloudwatch-container-insights/latest/k8s-deployment-manifest-templates/deployment-mode/daemonset/container-insights-monitoring/quickstart/cwagent-fluentd-quickstart.yaml | sed "s/{{cluster_name}}/dev/;s/{{region_name}}/us-west-1/" | kubectl apply -f -

Where dev is the name of the EKS cluser
us-west-1 is the region where cluster is deployed

When you execute the above command, it will create the amazon-cloudwatch namespace, set up the CloudWatch and Fluentd agent, create cloudwatch-agent service accounts, and create configmap for cloudwatch agent deploy cloudwatch agent as daemonsets.

serviceaccount/cloudwatch-agent created

clusterrole.rbac.authorization.k8s.io/cloudwatch-agent-role created

clusterrolebinding.rbac.authorization.k8s.io/cloudwatch-agent-role-binding created

configmap/cwagentconfig created

daemonset.apps/cloudwatch-agent created

configmap/cluster-info created

serviceaccount/fluentd created

clusterrole.rbac.authorization.k8s.io/fluentd-role created

clusterrolebinding.rbac.authorization.k8s.io/fluentd-role-binding created

configmap/fluentd-config created

daemonset.apps/fluentd-cloudwatch created

You can verify the status of CloudWatch and Fluentd pod using below commands

kubectl get pods -n amazon-cloudwatch

NAME                       READY   STATUS    RESTARTS   AGE

cloudwatch-agent-7rmlk     1/1     Running   0          4m29s

cloudwatch-agent-b9cth     1/1     Running   0          4m29s

cloudwatch-agent-vrgnl     1/1     Running   0          4m29s

fluentd-cloudwatch-29vxd   1/1     Running   0          4m28s

fluentd-cloudwatch-5rvqg   1/1     Running   0          4m28s

fluentd-cloudwatch-h52q4   1/1     Running   0          4m28s

NOTE: Attach CloudWatchAgentServerPolicy to the worker node before proceed further

Now go to the CloudWatch dashboard https://us-west-1.console.aws.amazon.com/cloudwatch , Under Insights, click on Container Insights

From the drop down select Performance monitoring and under Select clusters, select your EKS cluster.You will see the dashboard like this which display metrics like CPU, memory utilization and various Network statistics across EKS cluster.