As we’ve seen, integrating Kubernetes environments with Splunk Observability Cloud is a quick and easy way to get the end-to-end insight we need to keep our applications healthy and our customers happy. In this post, we’ll explore how to integrate Amazon Elastic Kubernetes Service (EKS) with Splunk Observability Cloud so we can observe EKS alongside the rest of our application telemetry data.
Observability Metrics in AWS
The Amazon EKS management console and Amazon CloudWatch provide insight into EKS observability metrics. From the AWS management console, you can see things like cluster health and details about EKS resources deployed to your cluster. Node details are available within a selected cluster’s Resources tab:
You can dig into pod status, capacity, and pod details:
Get an overview of node conditions like MemoryPressure, DiskPressure, etc.:
You can even get insight into node events like node status:
For more detailed cluster health visibility, you can enable CloudWatch Observability from the EKS console. With Container Insights enabled, you get even deeper insight into cluster state and resource utilization:
Along with cluster, namespace, node, service, workload, pod, and container performance monitoring:
But if your infrastructure isn’t all in AWS and is spread across multiple platforms, navigating between observability tools, especially during a high-pressure incident, is not great. Instead, having one unified observability platform where you can view all these metrics and more reduces toil and time to incident resolution. End-to-end visibility unified in a central platform makes for a more resilient and efficient observability practice. So let’s look at how to integrate one such observability platform, Splunk Observability Cloud, with Amazon EKS.
Integrate AWS and Splunk Observability Cloud
Splunk Observability Cloud provides a unified platform for troubleshooting and monitoring all application systems no matter where they live. Not only can you collect and store Amazon Cloudwatch Metrics data, but if pieces of your applications and infrastructure live outside of AWS, you can view that data right alongside your AWS data for a complete observability picture.
You may have already integrated AWS with Splunk Observability Cloud through the Data Management section in Splunk Observability Cloud:
The integration wizard easily takes you through the process of preparing your AWS account:
And getting your AWS data flowing into Splunk Observability Cloud:
But for EKS, data is collected using the Splunk Distribution of the OpenTelemetry Collector, and even with AWS integrated with Splunk Observability Cloud, you’ll notice from the Available Integrations page that we still need to deploy the OpenTelemetry Collector to get our EKS data in:
So let’s install the Splunk Distribution of the OpenTelemetry Collector for Kubernetes.
Install the Splunk Distribution of the OpenTelemetry Collector
We’ve gone through how to integrate Kubernetes and Splunk Observability Cloud before, and integrating Amazon EKS isn’t much different. When we follow along with the integration wizard, we just need to specify Amazon Web Services as the provider and Amazon EKS (or Amazon EKS / Fargate profiles) as the distribution:
We can connect to our EKS cluster and then follow along with the rest of the installation instructions. I’m using the AWS CLI in my terminal, but with Helm installed, you could also use AWS CloudShell. I first configured kubectl for my EKS cluster by updating my kubeconfig file:
And verified the connection:
I next ran the commands in the Splunk Observability Cloud installation instructions with splunk-otel-collector --version pinned to 0.111.0:
Once those steps were complete, I could then view my EKS telemetry data from within Splunk Observability Cloud:
From Infrastructure Monitoring we can view our Amazon EKS navigators:
We can get insight into all of our Kubernetes clusters:
Dive into the health of a specific cluster:
And observe critical performance data around nodes, containers, daemonsets, deployments, namespaces, pods, replicasets, and workloads:
From these critical usage metrics, we can create detectors and alerts from within our navigators and they can live right alongside the detectors and alerts for the rest of our applications and infrastructure:
With Amazon EKS now successfully integrated, we can use Splunk Observability Cloud to proactively monitor, detect, and alert on anomalies in our EKS environment right alongside the rest of our application and infrastructure telemetry data.
Wrap up
Integrating with a third-party observability platform like Splunk Observability Cloud provides a unified observability solution for your applications and infrastructure. This helps with quick and easy incident detection and resolution without having to navigate between a bunch of different observability solutions.
Want to try integrating Amazon EKS with Splunk Observability Cloud? Try Splunk Observability Cloud free for 14 days!
Resources
