Monitoring clusters and Pods using native AWS tools

One of the key advantages of an AWS deployment of Kubernetes (EKS) over an on-premises deployment of Kubernetes is it comes pre-integrated into CloudWatch, which is the main logging and monitoring platform for AWS. With a standard EKS cluster, you will automatically get control plane logs, EC2 worker node and load balancer (Network or Application Load Balancer) logs and metrics, along with metrics and logs from other AWS services such as databases, message queues, and so on.

Let’s look at how we can create a basic CloudWatch dashboard using standard EC2 metrics to understand the work nodes for our cluster.

Creating a basic CloudWatch dashboard

We’ll use Terraform to create a simple dashboard that shows an aggregated view of all instances that are tagged with a specific cluster name, and a second one that shows each individual node. The code snippet shown next illustrates the basic structure, a data object (which...