Hi LogicMonitor Community, We are trying to build a custom AWS_CLOUDWATCH DataSource to collect per-service CPU and memory metrics from AWS ECS Container Insights. Objective Our goal is to monitor CPUUtilized and MemoryUtilized for each ECS service in the prod-webapp cluster and display these metrics on a LogicMonitor dashboard. AWS Validation Completed We have verified using AWS CLI and CloudWatch that the following metrics are available and returning valid datapoints: Namespace: ECS/ContainerInsights Metrics: CPUUtilized , MemoryUtilized Dimensions: ClusterName = prod-cluser-app ServiceName = <ecs-service-name> Statistic: Average Custom DataSource Configuration Collection Method: AWS_CLOUDWATCH Active Discovery Method: CLOUDWATCH Namespace: ECS/ContainerInsights Metric Names: CPUUtilized, MemoryUtilized Device Dimension Name: ClusterName Instance Dimension Name: ServiceName Metrics Period: 900 Datapoint Metric Paths ECS/ContainerInsights>ClusterName:##system.aws.resourceid##>ServiceName:##wildvalue##>CPUUtilized>Average
ECS/ContainerInsights>ClusterName:##system.aws.resourceid##>ServiceName:##wildvalue##>MemoryUtilized>Average Current Issue The DataSource status shows Not in Use , even though the AWS metrics are confirmed to exist and return data through the AWS CLI. The standard AWS_ECS_ClusterServices DataSource is working correctly in our environment, so AWS credentials and ECS monitoring are functioning as expected. Question Has anyone successfully implemented a custom DataSource for ECS Container Insights metrics using AWS_CLOUDWATCH ? If so, are there any specific requirements or best practices for: AppliesTo expressions Active Discovery configuration CloudWatch metric path formatting Multi-instance setup Any guidance or examples would be greatly appreciated. Thank you!
