What problem does it solve? Setting up and operating AWS observability involves dozens of services with subtle constraints — alarm evaluation mechanics, EMF limits, X-Ray sampling rules, and ADOT auto-instrumentation steps — that are easy to misconfigure. This Skill provides domain expertise for building, debugging, and optimizing the full observability stack, including onboarding un-instrumented services to Application Signals. ## Core Features & Use Cases - Application Signals Onboarding: Enable ADOT auto-instrumentation on EC2, ECS, EKS, and Lambda for Python, Node.js, Java, and .NET, including ServiceEvents metadata, OTLP endpoints, and Dynamic Instrumentation. - CloudWatch Operations: Write Log Insights queries, configure metric/composite/anomaly alarms with best-practice defaults, publish custom metrics via EMF, and build dashboards. - Tracing & Auditing: Migrate from X-Ray SDK to ADOT, tune sampling rules, and perform CloudTrail operational auditing. - Use Case: A DevOps engineer needs to onboard a Python service on EKS to Application Signals and add error-rate alarms. The Skill guides the EKS add-on setup, IAM policies, inject annotations, and provides a CDK alarm template with p99 latency and M-of-N evaluation defaults. ## Quick Start Ask the assistant to enable Application Signals on your EKS Python service or to create a CloudWatch alarm for Lambda errors using best-practice defaults.