AWS recently announced the launch of Container Insights with Enhanced Observability for Amazon Elastic Container Service (ECS). This goes on the lines of a similar feature previously introduced for Amazon Elastic Kubernetes Service (EKS). This new capability aims to improve monitoring and troubleshooting for container workloads.
Donnie Prakoso, Principal Developer Advocate at AWS discussed the announcement in a blog post. The enhanced observability feature addresses a critical gap in container monitoring. It automates the collection of detailed performance metrics and logs. Users can now access granular metrics such as CPU utilization at both task and container levels without the need for extensive manual searches or deep application architecture knowledge. This automation allows for more efficient root-cause analysis and quicker identification of issues.
Users have two options for enabling this feature: cluster-level onboarding for specific clusters and account-level onboarding for all new clusters created within an account. To enable this feature, navigate to the Amazon ECS console, access account settings, and select the option for Container Insights with enhanced observability. This can be done during cluster creation or by updating existing clusters.
Once enabled, users can view task-level metrics in the cluster overview console. The Container Insights dashboard provides a honeycomb visualization that summarizes cluster health.
Source: Container Insights with enhanced observability now available in Amazon ECS
Furthermore, container insights with enhanced observabilty show granular resource usage patterns, that allow the users to quickly identify root causes by examining detailed resource usage patterns and correlating telemetry data. Curated dashboards based on AWS best practices enable proactive management of ECS resources. Deployment tracking allows users to track recent deployments alongside infrastructure anomalies, facilitating faster issue detection and rollback capabilities when necessary.
Cross-account monitoring provides a unified view of resources, easing operational overhead by supporting monitoring across multiple accounts. Integration with other CloudWatch services like Application Signals and CloudWatch Logs enables users to correlate infrastructure performance with running services.
Container Insights feature employs a dual-state monitoring approach: an alarm state indicating whether specific thresholds are met (red or green) and a utilization state using color coding (dark blue or light blue) to reflect resource usage patterns. Users can hover over clusters to see alarms at different layers, from the cluster down to the container level. A list format is also available for cross-account observability, displaying account IDs and labels for quick identification of ownership.
The detailed dashboard allows users to investigate spikes in resource utilization further. Filters are available for conducting thorough investigations across containers, services, or tasks. Users can access various logs—structured performance logs and application/container logs—to trace events leading to issues effectively.
Recently, AWS was in the news as they made two announcements. First, AWS Data Transfer Terminal, a new service currently available in the US, that provides high-speed data uploads through dedicated physical locations. Customers can bring their storage devices to these facilities, which provides at least two 100G optical fiber connections to the AWS network, for fast data transfer to and from the AWS cloud.
Second, AWS Asia Pacific (Thailand) Region marks AWS’s entry into the Thai market with its first infrastructure region in the country and its fourteenth in the Asia Pacific region.
Additionally, container insights’ integration with AWS X-Ray traces helps users analyze application behavior within containers. This enhanced capability also works with Amazon CloudWatch Application Signals, providing visibility into application health and performance against service-level objectives.
Container Insights with enhanced observability is now available in all AWS regions, including China. The pricing model features flat metric pricing. To learn more about improved container insights, head over to the Amazon CloudWatch documentation.
