Amazon Elastic Kubernetes Service (EKS) is a cloud-based compute platform that includes a fully managed Kubernetes control plane in order to simplify cluster operations. AWS introduced EKS Anywhere to bring the operational ease of EKS to organizations that manage on-premise environments (e.g., to meet data sovereignty requirements). EKS Anywhere is released as an automation tool that launches an EKS Distro cluster with opinionated defaults on vSphere virtual machines, making it easier to get started with container orchestration.
With EKS Anywhere, teams can now use consistent tooling to operate and manage both cloud-based and on-premise EKS clusters. This makes EKS Anywhere a good fit for organizations that plan to migrate their on-premise Kubernetes deployments to the cloud, as well as organizations that usually run workloads on premises but want to deploy to AWS in order to handle bursts in traffic.
We are proud to announce that Datadog is a launch partner for EKS Anywhere. Datadog enables you to get full visibility into the health and performance of EKS Anywhere and cloud-based EKS workloads—and their underlying infrastructure, whether it’s running on premises or in the cloud.
Whether you deploy your EKS clusters on your own data centers, to the cloud, or both, you need to monitor the resource usage and availability of both environments in order to keep your applications running as expected. With Datadog’s Kubernetes resources view (part of Live Containers), you can get real-time insights into the resource capacity and usage of your EKS and EKS Anywhere clusters, all in one platform.
In the example below, we are viewing pods from two EKS clusters: one hosted on premises (
prod-11287-demo-cluster-west) and the other in the AWS cloud. While this view shows us that CPU and memory utilization for pods in each cluster are similar, if we run into capacity issues in our on-premise data centers, we can scale out our cloud-based EKS cluster until the issue subsides.
While EKS Anywhere makes it easier to manage your on-premise Kubernetes deployments, you will still need to monitor your underlying infrastructure to prevent issues with scheduling your application pods. Datadog enables you to explore bespoke views of your EKS Anywhere deployments—and get full context around the vSphere VMs that host them—so you can investigate unavailable nodes, resource saturation, and other conditions that can stop EKS Anywhere from working as expected.
With Datadog, you can create custom dashboards to monitor Kubernetes metrics, such as the counts of available and desired pods, as well as vSphere metrics. For example, you can create a dashboard (as shown below) that graphs the number of vSphere VMs by regional data center, then graph the number of available Kubernetes nodes by region. If there are more VMs than Kubernetes nodes, you can investigate whether some kubelets running on your vSphere VMs have crashed—or failed to deploy due to a misconfiguration—causing the API server to register fewer nodes than expected.
Datadog is well suited for monitoring EKS Anywhere as well as your cloud-based EKS deployments, giving you full visibility into your containerized workloads no matter where they run. You can quickly install Datadog in your EKS Anywhere and EKS clusters using our Helm chart.
Datadog integrates with vSphere—and other infrastructure technologies you might be deploying on Kubernetes, such as Cilium and CoreDNS—so you can visualize every layer of your EKS Anywhere environment. And with other features like Network Device Monitoring, you can get even deeper insight into the health of your on-premise infrastructure.
If you have not yet signed up for Datadog, you can get started today with a free trial.