AWS Improves Container Monitoring, Part 2: Container Insights
AWS is previewing Container Insights, a CloudWatch component that allows for greater visibility into containerized applications using microservice architectures.
The goal is to provide IT operations teams with improved observability of their applications. Observability comes from a combination of people, process, and technology.
CloudWatch Container Insights aims to beef up the technology component for improved observability. Let’s dive into some details.
How It Works
- Users install the Container Insights agent on the cluster.
- Clusters supported include EKS, ECS, and Fargate.
- After installation, Container Insights will appear on the CloudWatch dropdown in the AWS console.
User Interface
- Users can select the cluster view and can filter between different container clusters.
- The console presents a dashboard that shows resource utilization for each cluster: compute, memory, storage, and network.
- Users can also see the nodes and the node health, any alarms that are attached to the nodes, and their state (i.e. whether the alarms have been triggered or not).
- Users are also able to view metrics from a namespace perspective (i.e. not just from a cluster perspective). This is useful because many teams use name spaces to distinguish between applications.
Logs
- Users can pull logs from the cluster and can perform analytics on those logs.
- The console will present a log group option – this is a grouping of related logs.
- Users can select multiple log groups if they wish to perform correlation across different log groups.
- Users can also drill down into the host-level logs.
- A dashboard graph shows the distribution of logs over time.
- The console allows for aggregation and parsing of logs to display on a dashboard.
- AWS recommends JSON logs for this purpose, because then the console will auto-discover the fields (however, users can still parse XML files for the fields).
- Users can also view the query history for the account and can rerun previous queries.
We can see that Container Insights provides users with a fairly robust monitoring dashboard out of the box.
Because Container Insights is still in preview, AWS recommends limiting its use to test environments and not using it in production until after the general availability release.
Our Take
Containers can offer great advantages for development and for infrastructure as code capabilities, but they can create a lot of extra work for operations. Container Insights helps bridge this gap by providing powerful out-of-the-box monitoring capabilities to operations teams.
See the other notes in this series for more detail about container observability and Anomaly Detection.
Want to Know More?
AWS Improves Container Monitoring, Part 1: Observability
AWS Improves Container Monitoring, Part 3: Anomaly Detection