Optimize AWS costs and monitor cloud operations

Comprehensive AWS cost optimization, monitoring, observability, and operational-excellence guidance built around three bundled MCP servers: AWS Pricing (pricing, for pre-deployment cost estimation, regional pricing comparison, and TCO calculation), AWS Cost Explorer (costexp, for historical spending analysis, anomaly detection, and forecasting), and Amazon CloudWatch (cw, for metrics, alarms, and log queries plus resource-utilization monitoring), plus five additional MCP servers - Billing and Cost Management, Application Signals, Managed Prometheus, CloudTrail, and Well-Architected Security Assessment - available separately via a companion aws-mcp-setup skill's Full AWS MCP Server. It requires verifying AWS facts through the actual MCP tools before answering, auto-loading the aws-mcp-setup dependency and guiding the user through its setup flow if the tools aren't available. It documents three end-to-end workflows: a cost-analysis workflow (pre-deployment estimation via Pricing MCP, post-deployment tracking via Billing MCP, detailed analysis via Cost Explorer, then acting on its recommendations), a monitoring workflow (CloudWatch alarm setup, tracking key metrics, Application Signals for APM, CloudWatch Logs Insights for troubleshooting), and a security workflow (CloudTrail activity review, Well-Architected Security Assessment, remediation, ongoing monitoring). It gives concrete cost-optimization strategies - right-sizing over-provisioned resources, appropriate S3/EBS storage classes, auto-scaling, Savings Plans and Reserved Instances, deleting unused resources and snapshots, consistent cost-allocation tags - named alarm scenarios such as Lambda error rate over 1%, EC2 CPU over 80%, API Gateway 4xx/5xx spikes, DynamoDB throttling, and ECS task failures, and named CloudTrail audit queries like who deleted a bucket, IAM role changes in the last 24 hours, failed logins, actions by a specific user, and security-group modifications. Operational-excellence guidelines cover tagging everything, monthly spend reviews, right-sizing, automation, budget alerts, noise-reduced and actionable monitoring alerts with clear remediation steps, dashboard visibility for stakeholders, least-privilege access, regular CloudTrail audits, encryption at rest and in transit, and having an incident-response procedure. Two bundled reference files, references/operations-patterns.md and references/cloudwatch-alarms.md, hold deeper operational patterns and per-service alarm configurations for Lambda, EC2, RDS, DynamoDB, API Gateway, ECS, and Application Load Balancers.

Use it when optimizing AWS costs, estimating costs before deployment, monitoring application or infrastructure performance, setting up observability and alerting, analyzing spending trends, investigating operational issues, auditing AWS activity, assessing security posture, or implementing operational excellence broadly.

Input is an AWS cost, monitoring, or security question, or a resource configuration to estimate or analyze. Output is an MCP-tool-verified answer - a cost estimate, a spending analysis, a CloudWatch alarm or dashboard configuration, or an audit or security finding - for example a pre-deployment estimation query:

"Estimate the monthly cost of running a Lambda function with
1 million invocations, 512MB memory, 3-second duration in us-east-1"

Built on three bundled MCP servers - AWS Pricing, Cost Explorer, and CloudWatch - with five more available via the aws-mcp-setup skill's Full AWS MCP Server (Billing and Cost Management, Application Signals, Managed Prometheus, CloudTrail, Well-Architected Security Assessment), covering Lambda, EC2, RDS, DynamoDB, API Gateway, ECS, and Application Load Balancers.

Cloud engineers and FinOps practitioners who need MCP-verified AWS cost estimation, spend analysis, monitoring, and security-audit workflows rather than answers from unverified model knowledge of AWS pricing and services.