ecs_skill

Overview

This skill provides practical AWS ECS operations for deploying and managing Docker containers on Fargate or EC2. It guides you through creating clusters, registering task definitions, running tasks, and configuring services with load balancers and autoscaling. The content focuses on commands, patterns, and troubleshooting steps you can apply immediately.

How this skill works

The skill inspects and automates common ECS workflows: cluster creation, task definition registration, service creation and updates, standalone task runs, and autoscaling configuration. It includes concrete AWS CLI examples and diagnostic commands to inspect stopped tasks, logs, deployments, and service events. Use the patterns to deploy applications, update images, and resolve common failures quickly.

When to use it

• Deploy containerized applications to AWS using ECS (Fargate or EC2).
• Define and register task definitions with containers, environment variables, secrets, and logging.
• Create or update services with load balancers and health checks.
• Run batch or one-off jobs as standalone tasks.
• Troubleshoot tasks that fail to start, containers that keep restarting, or services stuck deploying.

Best practices

• Use IAM task roles for application permissions and execution roles for ECR and secrets access.
• Store sensitive values in Secrets Manager or Parameter Store; avoid embedding secrets in images.
• Right-size CPU and memory per task and enable container insights for monitoring.
• Deploy across multiple Availability Zones and configure health checks and deregistration delays.
• Use capacity providers and Fargate Spot for cost optimization and scalable capacity.

Example use cases

• Create a Fargate cluster and register a web-app task definition with awslogs logging and health checks.
• Deploy a two-task web service behind an ALB target group and enable a deployment circuit breaker for safer rollbacks.
• Run a batch job as a standalone Fargate task with a private subnet and assignPublicIp as needed.
• Update a service by registering a new task definition with an updated image and force a new deployment.
• Configure target-tracking autoscaling for an ECS service based on average CPU utilization.