Once you’ve integrated with AWS CloudWatch, you have access to metrics from AWS ECS, which is an incredibly quick and scalable container management service. It can be used to manage, run, and stop containers on a cluster. Your containers using Amazon ECS are specified in a task description that you use to execute a single task or a task inside of a service.
See all AWS integrations.
To verify metrics are reporting, search for the metrics on the Metric details page in Settings.
The following table shows the ECS metrics ingested by Cloud Observability.
| Metric Name | Unit | Description |
|---|---|---|
| aws.ecs.cpu_reservation | gauge | The portion of the cluster's CPUs that are set aside for active tasks. The total CPU units that are reserved by Amazon ECS tasks on the cluster are divided by the total CPU units that were registered for all of the container instances in the cluster to calculate cluster CPU reservation (this metric may only be filtered by ClusterName). CPU reservation measurements only apply to container instances with an ACTIVE or DRAINING status. Only tasks employing the EC2 launch type use this metric. |
| aws.ecs.cpu_utilization | gauge | The proportion of CPUs being used by the cluster or service. Cluster CPU utilization is calculated as the total number of CPUs used by Amazon ECS jobs on the cluster divided by the total number of CPUs registered for all of the container instances in the cluster (metrics that are filtered by ClusterName without ServiceName). Only instances of containers with an ACTIVE or DRAINING status will have an impact on CPU utilization metrics. Only processes employing the EC2 launch type use cluster CPU utilization metrics. Service When metrics are filtered by ClusterName and ServiceName, CPU utilization is calculated as the sum of CPU units used by tasks associated with the service divided by the sum of CPU units reserved for. |
| aws.ecs.memory_reservation | gauge | The portion of RAM that is set aside for active cluster processes. Cluster memory reservation is calculated as the total amount of memory that is reserved by Amazon ECS tasks on the cluster, divided by the total amount of memory that was registered for all of the container instances in the cluster (this metric may only be filtered by ClusterName). Memory reservation measurements only apply to container instances with an ACTIVE or DRAINING status. Only tasks employing the EC2 launch type use this metric. |
| aws.ecs.memory_utilization | gauge | The amount of memory that the cluster or service is currently using. Cluster memory usage is calculated as the total amount of memory used by Amazon ECS jobs on the cluster divided by the total amount of memory that was registered for all of the container instances in the cluster (metrics that are filtered by ClusterName without ServiceName). Metrics for memory use are only affected by container instances with an ACTIVE or DRAINING status. Metrics for cluster memory usage are only used for jobs that employ the EC2 launch type. The amount of memory used by the tasks associated with a service, divided by the amount of memory allotted for those jobs, is the measure of service memory usage (metrics that are filtered by ClusterName and ServiceName). |
| aws.ecs.gpu_reservation | gauge | The portion of the cluster's total GPU capacity that is used by active jobs. The amount of GPUs reserved by Amazon ECS tasks on the cluster is calculated as the total number of GPUs available on all of the container instances with GPUs in the cluster, divided by the number of GPUs that were reserved by Amazon ECS tasks on the cluster. Only container instances with an ACTIVE or DRAINING status will have an impact on the metrics for GPU reservations. |
See also
Updated Dec 1, 2022