What problem does it solve? When ECS tasks stop sending logs to CloudWatch, it is hard to tell whether the failure is cluster-wide or limited to specific nodes. This Skill pinpoints Docker awslogs failures on EC2-backed ECS nodes by inspecting Docker journal errors and building a failing-versus-clean node matrix. ## Core Features & Use Cases - Per-Node Journal Probing: Run SSM commands against one or more EC2 instances to grep the Docker journal for awslogs errors such as failed log stream creation and credential refresh failures. - Cluster-Wide Node Matrix: Automatically list active ECS container instances, map them to EC2 IDs, and probe each node to produce a failing-versus-clean comparison saved to a durable temp path. - Known Failure Pattern Reference: Consult documented guidance on the awslogs credential-refresh failure mode, including why drain and replacement is safer than assuming a Docker restart fixes it. - Use Case: An on-call engineer notices missing CloudWatch logs for some ECS services. They check one suspect node, compare it against a clean node, then build a cluster-wide matrix to identify the full failing-node set before draining. ## Quick Start Ask the agent to check whether ECS node i-0123456789abcdef0 has Docker awslogs failures since yesterday using your AWS profile and region.
