CFM Tips - Cost Optimization MCP Server

A comprehensive Model Context Protocol (MCP) server for AWS cost analysis and optimization recommendations, designed to work seamlessly with Amazon Q CLI and other MCP-compatible clients.

🚀 Quick Start

# Clone the repository
git clone https://github.com/aws-samples/sample-cfm-tips-mcp.git

cd sample-cfm-tips-mcp

# Install dependencies
pip install -r requirements.txt

# Configure AWS credentials
aws configure

# Add the MCP server config to Amazon Q using the mcp_runbooks.json as a template
vi ~/.aws/amazonq/mcp.json

# Example usage in Q chat:
"Run comprehensive cost analysis for us-east-1"
"Find unused EBS volumes costing money"
"Generate EC2 right-sizing report in markdown"

🧩 Add-on MCPs

Add-on AWS Pricing MCP Server MCP server for accessing real-time AWS pricing information and providing cost analysis capabilitieshttps://github.com/awslabs/mcp/tree/main/src/aws-pricing-mcp-server

# Example usage with Add-on AWS Pricing MCP Server:
"Review the CDK by comparing it to the actual spend from my AWS account's stackset. Suggest cost optimization opportunities for the app accordingly"

✅ Features

Core AWS Services Integration

Cost Explorer - Retrieve cost data and usage metrics
Cost Optimization Hub - Get AWS cost optimization recommendations
Compute Optimizer - Right-sizing recommendations for compute resources
Trusted Advisor - Cost optimization checks and recommendations
Performance Insights - RDS performance metrics and analysis
CUR Reports - Cost and Usage Report analysis from S3

Cost Optimization Playbooks

🔧 EC2 Right Sizing - Identify underutilized EC2 instances
💾 EBS Optimization - Find unused and underutilized volumes
🗄️ RDS Optimization - Identify idle and underutilized databases
⚡ Lambda Optimization - Find overprovisioned and unused functions
📊 Comprehensive Analysis - Multi-service cost analysis

Advanced Features

Real CloudWatch Metrics - Uses actual AWS metrics for analysis
Multiple Output Formats - JSON and Markdown report generation
Cost Calculations - Estimated savings and cost breakdowns
Actionable Recommendations - Priority-based optimization suggestions

📁 Project Structure

sample-cfm-tips-mcp/
├── playbooks/                            # CFM Tips optimization playbooks engine
├── services/                             # AWS Services as datasources for the cost optimization
├── mcp_server_with_runbooks.py           # Main MCP server
├── runbook_functions.py                  # Cost optimization runbook implementations
├── mcp_runbooks.json                     # Template file for MCP configuration file
├── requirements.txt                      # Python dependencies
├── test_runbooks.py                      # Integration tests
├── diagnose_cost_optimization_hub_v2.py  # Diagnostic utilities
├── RUNBOOKS_GUIDE.md                     # Detailed usage guide
└── README.md                             # Project ReadMe

🔐 Required AWS Permissions

Your AWS credentials need these permissions. The MCP server uses your role to perform actions, hence proceed with least privilege access only. The below creates an IAM policy with for list, read and describe actions only:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "cost-optimization-hub:ListEnrollmentStatuses",
        "cost-optimization-hub:ListRecommendations",
        "cost-optimization-hub:GetRecommendation", 
        "cost-optimization-hub:ListRecommendationSummaries",
        "ce:GetCostAndUsage",
        "ce:GetCostForecast",
        "compute-optimizer:GetEC2InstanceRecommendations",
        "compute-optimizer:GetEBSVolumeRecommendations",
        "compute-optimizer:GetLambdaFunctionRecommendations",
        "ec2:DescribeInstances",
        "ec2:DescribeVolumes",
        "rds:DescribeDBInstances",
        "lambda:ListFunctions",
        "cloudwatch:GetMetricStatistics",
        "s3:ListObjectsV2",
        "support:DescribeTrustedAdvisorChecks",
        "support:DescribeTrustedAdvisorCheckResult",
        "pi:GetResourceMetrics"
      ],
      "Resource": "*"
    }
  ]
}

🛠️ Installation

Prerequisites

Python 3.11 or higher
AWS CLI configured with appropriate credentials
Amazon Q CLI (for MCP integration) - https://docs.aws.amazon.com/amazonq/latest/qdeveloper-ug/command-line-installing.html

Setup Steps

Clone the Repository

git clone https://github.com/aws-samples/sample-cfm-tips-mcp.git
cd sample-cfm-tips-mcp

Install Dependencies
```
pip install -r requirements.txt
```

Configure AWS Credentials

aws configure
# Or set environment variables:
# export AWS_ACCESS_KEY_ID=your_access_key
# export AWS_SECRET_ACCESS_KEY=your_secret_key
# export AWS_DEFAULT_REGION=us-east-1

Apply IAM Permissions
- Create an IAM policy with the permissions listed above
- Attach the policy to your IAM user or role
Test the Installation
```
python3 test_runbooks.py
```
Start the MCP Server
```
python3 mcp_server_with_runbooks.py &
```
Start the Q Chat
```
q chat
```

🔧 Available Tools

Cost Analysis Tools

get_cost_explorer_data - Retrieve AWS cost and usage data
list_coh_enrollment - Check Cost Optimization Hub enrollment
get_coh_recommendations - Get cost optimization recommendations
get_coh_summaries - Get recommendation summaries
get_compute_optimizer_recommendations - Get compute optimization recommendations

EC2 Optimization

ec2_rightsizing - Analyze EC2 instances for right-sizing opportunities
ec2_report - Generate detailed EC2 optimization reports

EBS Optimization

ebs_optimization - Analyze EBS volumes for optimization
ebs_unused - Identify unused EBS volumes
ebs_report - Generate EBS optimization reports

RDS Optimization

rds_optimization - Analyze RDS instances for optimization
rds_idle - Identify idle RDS instances
rds_report - Generate RDS optimization reports

Lambda Optimization

lambda_optimization - Analyze Lambda functions for optimization
lambda_unused - Identify unused Lambda functions
lambda_report - Generate Lambda optimization reports

Comprehensive Analysis

comprehensive_analysis - Multi-service cost analysis

Additional Tools

list_cur_reports - List Cost and Usage Reports in S3
get_trusted_advisor_checks - Get Trusted Advisor recommendations
get_performance_insights_metrics - Get RDS Performance Insights data

📊 Example Usage

Basic Cost Analysis

"Get my AWS costs for the last month"
"Show me cost optimization recommendations"
"What are my biggest cost drivers?"

Resource Optimization

"Find underutilized EC2 instances in us-east-1"
"Show me unused EBS volumes that I can delete"
"Identify idle RDS databases"
"Find unused Lambda functions"

Report Generation

"Generate a comprehensive cost optimization report"
"Create an EC2 right-sizing report in markdown format"
"Generate an EBS optimization report with cost savings"

Multi-Service Analysis

"Run comprehensive cost analysis for all services in us-east-1"
"Analyze my AWS infrastructure for cost optimization opportunities"
"Show me immediate cost savings opportunities"

🔍 Troubleshooting

Common Issues

Cost Optimization Hub Not Working

python3 diagnose_cost_optimization_hub_v2.py

No Metrics Found
- Ensure resources have been running for at least 14 days
- Verify CloudWatch metrics are enabled
- Check that you're analyzing the correct region
Permission Errors
- Verify IAM permissions are correctly applied
- Check AWS credentials configuration
- Ensure Cost Optimization Hub is enabled in AWS Console

Import Errors

# Check Python path and dependencies
python3 -c "import boto3, mcp; print('Dependencies OK')"

Getting Help

Check the RUNBOOKS_GUIDE.md for detailed usage instructions
Run the diagnostic script: python3 diagnose_cost_optimization_hub_v2.py
Run integration tests: python3 test_runbooks.py

🎯 Key Benefits

Immediate Cost Savings - Identify unused resources for deletion
Right-Sizing Opportunities - Optimize overprovisioned resources
Real Metrics Analysis - Uses actual CloudWatch data
Actionable Reports - Clear recommendations with cost estimates
Comprehensive Coverage - Analyze EC2, EBS, RDS, Lambda, and more
Easy Integration - Works seamlessly with Amazon Q CLI

📈 Expected Results

The CFM Tips cost optimization server can help you:

Identify cost savings on average
Find unused resources costing hundreds of dollars monthly
Right-size overprovisioned instances for optimal performance/cost ratio
Optimize storage costs through volume type recommendations
Eliminate idle resources that provide no business value

🤝 Contributing

We welcome contributions! Please see our contributing guidelines:

Fork the repository
Create a feature branch
Make your changes
Add tests for new functionality
Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.