Amazon CloudWatch Container Insights for Amazon EKS: A Comprehensive Guide

/ Tutorial

In April 2026, Amazon CloudWatch significantly enhanced its monitoring capabilities with the introduction of OTel (OpenTelemetry) Container Insights for Amazon EKS (Elastic Kubernetes Service). This feature is designed for developers and operators aiming to gain critical visibility into their Kubernetes clusters, helping them manage containers more effectively. In this comprehensive guide, we will explore the new functionality of Amazon CloudWatch, touch on the technical details, provide actionable insights, and equip you with everything you need to maximize the potential of OTel Container Insights for your Amazon EKS deployments.

Table of Contents

  1. Introduction to Amazon CloudWatch and EKS
  2. What is OpenTelemetry and How Does it Work?
  3. Key Features of OTel Container Insights
  4. Setting Up OTel Container Insights
  5. 4.1 Using the Amazon EKS Console
  6. 4.2 Deploying via CloudFormation, CDK, or Terraform
  7. Collecting Metrics with OTel Container Insights
  8. 5.1 Understanding Metrics Collection
  9. 5.2 Enrichment of Metrics with Descriptive Labels
  10. Visualizing Metrics in CloudWatch
  11. 6.1 Using Curated Dashboards
  12. 6.2 Aggregating and Filtering Metrics
  13. Querying with Prometheus Query Language (PromQL)
  14. Advanced Features with the Observability EKS Add-on
  15. Practical Use Cases for OTel Container Insights
  16. Best Practices for Monitoring EKS Clusters
  17. Conclusion and Future Considerations

1. Introduction to Amazon CloudWatch and EKS

Amazon Elastic Kubernetes Service (EKS) allows developers to run Kubernetes applications in the cloud. Managed by AWS, it simplifies the complexity associated with the installation and maintenance of your cluster. To ensure the smooth operation of these applications, effective monitoring is critical. That’s where Amazon CloudWatch comes in.

With the launch of OTel Container Insights for Amazon EKS, CloudWatch not only broadens its monitoring capabilities but also integrates OpenTelemetry, an industry-standard for observability. This new addition brings advanced metrics collection, richer context, and powerful visualization capabilities that fundamentally change how developers interact with their Kubernetes deployments.

2. What is OpenTelemetry and How Does it Work?

OpenTelemetry is a collection of tools, APIs, and SDKs that can be used to instrument your applications for distributed tracing and metrics collection. It is designed to facilitate observability by providing developers with the means to collect telemetry data from their applications—regardless of the programming language or framework being used.

How OpenTelemetry Works:

  • Instrumentation: With OpenTelemetry, you can instrument your code to collect metrics and traces without being tied to a specific vendor.
  • Data Collection: OpenTelemetry supports numerous data sources, including application logs, Kubernetes events, and performance metrics.
  • Data Exporting: Once collected, data is sent to a backend system, such as Amazon CloudWatch, for monitoring and visualization.

By adopting OpenTelemetry, Amazon CloudWatch users can seamlessly collect a wider variety of metrics while enriching their monitoring data, ultimately enhancing their observability strategy.

3. Key Features of OTel Container Insights

The key features of OTel Container Insights enhance the existing Container Insights experience, bringing a wealth of functionality to EKS cluster monitoring:

  • Enhanced Metric Collection: Collects extensive metrics using both AWS services and open-source components.
  • Automatic Metric Enrichment: Metrics are auto-enriched with up to 150 labels, including Kubernetes metadata like pod names, deployment strategies, and environment tags.
  • Curated Dashboards: Ready-made dashboards provide at-a-glance health summaries of your EKS clusters, nodes, and pods.
  • Advanced Querying: Built-in support for Prometheus Query Language (PromQL) allows users to perform in-depth data analysis.
  • One-Click Installation: Easily deploy the Observability EKS add-on via the Amazon EKS console or infrastructure as code tools like Terraform and CloudFormation.

Together, these features position OTel Container Insights as a powerful tool for developers and operators looking to obtain greater insights into their containerized environments.

4. Setting Up OTel Container Insights

Adopting OTel Container Insights for your EKS environment involves a few straightforward steps. In this section, we will guide you through the installation and initial configuration.

4.1 Using the Amazon EKS Console

  1. Log into the AWS Management Console.
  2. Navigate to the Amazon EKS section.
  3. Choose the cluster where you want to enable OTel Container Insights.
  4. Select the Monitoring tab and enable Container Insights with OpenTelemetry.
  5. Confirm your settings and wait for the installation to complete.

4.2 Deploying via CloudFormation, CDK, or Terraform

You may prefer using Infrastructure as Code (IaC) tools for more flexibility or automate deployments:

  • For CloudFormation:
  • Prepare a CloudFormation template that provisions necessary IAM roles and K8s components.

  • Deploy the stack in your desired region.

  • For CDK:

  • Write a TypeScript or Python script that defines your EKS cluster alongside the OTel monitoring configuration.

  • Execute cdk deploy to provision the entire setup.

  • For Terraform:

  • Create a .tf file that includes the cluster definition and OTel settings.
  • Run terraform apply to deploy your resources.

Each of these approaches enables a streamlined installation process, fostering scalability and version control for your monitoring setup.

5. Collecting Metrics with OTel Container Insights

Understanding the metrics collected by OTel Container Insights is vital for effective monitoring and troubleshooting.

5.1 Understanding Metrics Collection

OTel Container Insights automates the collection of essential metrics such as:

  • CPU Utilization: Percentage of container usage versus total allocated.
  • Memory Utilization: Metrics to gauge memory consumption per pod or node.
  • Network Traffic: Incoming and outgoing network traffic for your containers.

You can access these metrics directly in the CloudWatch console and utilize them for real-time monitoring or historical analysis.

5.2 Enrichment of Metrics with Descriptive Labels

OTel automatically enriches each collected metric with valuable context that helps identify issues faster. Key labels include:

  • Kubernetes Metadata: Information regarding the object types (pods, deployments, nodes) in the cluster.
  • Custom Labels: User-defined labels like the name of the application, ownership team, and environment (production, staging, etc.).

This label enrichment plays a crucial role in filtering and grouping your metrics efficiently, enabling rapid identification of problems.

6. Visualizing Metrics in CloudWatch

Visualization is an integral part of monitoring, and CloudWatch provides robust visualization tools.

6.1 Using Curated Dashboards

The default curated dashboards in CloudWatch for Container Insights display crucial metrics, allowing you to quickly assess the health of your resources.

  • Cluster Health: See overall cluster status, node status, and resource allocation.
  • Node Metrics: View CPU, memory utilization, and disk I/O metrics for each node.
  • Pod Metrics: Gather insights on pod statuses and their performance metrics.

These dashboards are pre-built, saving time on the setup and providing immediate insights.

6.2 Aggregating and Filtering Metrics

With CloudWatch, you can further drill down by aggregating and filtering metrics based on:

  • Instance Types: Group metrics by instance types for analysis of performance across different workloads.
  • Custom Labels: Use your defined labels to parse metrics by application or environment, enable comparative analyses, and pinpoint areas of concern.

Creating customized views in CloudWatch enhances your monitoring capabilities and supports proactive management of your EKS clusters.

7. Querying with Prometheus Query Language (PromQL)

When in-depth analysis is required, the integration of Prometheus Query Language (PromQL) unlocks advanced querying capabilities within CloudWatch.

Why Use PromQL?

  • Complex Queries: Write more complex queries that delve deep into metrics beyond simple aggregations.
  • Time Series Analysis: Analyze time series data efficiently, allowing you to track changes in key performance indicators over time.
  • Filtering and Label Matching: Use labels to filter queries for specific instances or components, honing in on particular applications or resources.

Becoming proficient in PromQL can dramatically improve your ability to troubleshoot and analyze performance metrics.

8. Advanced Features with the Observability EKS Add-on

With the Observability EKS add-on, CloudWatch embraces modern infrastructure by supporting important features:

  • Auto-Detection of Hardware: Detect specialized hardware like NVIDIA GPUs and AWS Trainium/AWS Inferentia accelerators.
  • Multiple Metric Types: Publish both OTel metrics and existing Container Insights metrics simultaneously, allowing you to retain legacy monitoring as you transition to OTel.

These features make managing advanced workloads easier, providing performance insights that were previously challenging to obtain.

9. Practical Use Cases for OTel Container Insights

OTel Container Insights is not just about monitoring; it enables various use cases that enhance operations and management:

  • Performance Optimization: Identify underutilized resources and scale them to minimize costs.
  • Troubleshooting: Pinpoint and resolve performance bottlenecks based on enriched metrics and real-time dashboards.
  • Capacity Planning: Analyze long-term trends using historical data, allowing for informed decision-making in resource provisioning.
  • Security Monitoring: Assess application performance while ensuring compliance through monitoring suspicious activities.

By leveraging OTel Container Insights, organizations can operate more effectively and efficiently.

10. Best Practices for Monitoring EKS Clusters

To maximize the benefits of OTel Container Insights, adhere to these best practices:

  1. Define Custom Metrics: Create custom metrics that align with your specific business needs for more meaningful insights.
  2. Regular Query Testing: Continuously test and refine your queries to ensure you’re capturing the most relevant data.
  3. Manage Metric Retention: Set appropriate retention policies for metrics data within CloudWatch to balance cost and data availability.
  4. Automate Deployments: Use IaC for consistent and reliable deployment of your monitoring setup.
  5. Ongoing Training: Ensure team members are well-trained in the use of AWS services and Prometheus for maximum support and troubleshooting efficiency.

By observing these practices, you can ensure that your monitoring is robust and adaptable to changing business needs.

11. Conclusion and Future Considerations

Amazon CloudWatch’s launch of OTel Container Insights for Amazon EKS represents a significant leap forward in the observability landscape. By employing OpenTelemetry, users can achieve deeper visibility into their Kubernetes environments through comprehensive metric collection, enriched data, and insightful visualizations.

As cloud technologies evolve, staying ahead of trends in observability will be crucial for organizations that aim to scale efficiently. The integration of advanced querying capabilities and support for specialized hardware means that monitoring will only become more sophisticated in the future.

To leverage these benefits fully, familiarize yourself not only with OTel Container Insights but also with continuity in training your team on best practices for monitoring EKS clusters.

Key Takeaway: Implementing OTel Container Insights for your Amazon EKS environment is a strategic step towards achieving unparalleled observability and operational excellence.

Explore the power of Amazon CloudWatch with OTel Container Insights for Amazon EKS today!

Learn more

More on Stackpioneers

Other Tutorials