CloudWatch MCP Server
This simplified MCP server provides a streamlined way to interact with AWS CloudWatch resources through the MCP protocol. It exposes CloudWatch log groups, log queries, and alarms as resources and tools.
Features
List all CloudWatch log groups with their metadata
List all CloudWatch alarms with their current states
Query CloudWatch logs using CloudWatch Insights across multiple log groups
Discover available fields across multiple log groups with shared schema
Automatic JSON parsing for @message field in log queries
Check if specific log groups exist
Get detailed information about specific log groups
Filter alarms by state (all alarms or only those in ALARM state)
Retrieve all saved CloudWatch Logs Insights queries
Prerequisites
Python 3.12 or higher
AWS credentials configured (via environment variables, AWS CLI, or IAM role)
MCP CLI (version 0.1.1 or higher)
Boto3 (AWS SDK for Python)
Setup
Make sure you have Python 3.12+ installed.
Create a virtual environment (optional but recommended):
python -m venv .venv source .venv/bin/activate # On Windows: .venv\Scripts\activateInstall dependencies:
pip install -r requirements.txtConfigure AWS credentials if you haven't already:
aws configureOr set environment variables:
export AWS_ACCESS_KEY_ID="your-access-key" export AWS_SECRET_ACCESS_KEY="your-secret-key" export AWS_REGION="your-region"
Project Structure
cloudwatch_server.py
- MCP server implementation for CloudWatch integrationaws_cloudwatch.py
- Simplified AWS CloudWatch integration moduletest_cloudwatch.py
- Command-line utility to test the CloudWatch integration
Running the server
Start the MCP server:
Or using the MCP CLI:
Using the MCP server
Resources
The server exposes the following resources:
cloudwatch://log-groups
- Lists all CloudWatch log groupscloudwatch://log-groups/{log_group_name}
- Gets detailed information about a specific log groupcloudwatch://alarms
- Lists all CloudWatch alarmscloudwatch://alarms/in-alarm
- Lists only CloudWatch alarms currently in ALARM statecloudwatch://saved-queries
- Lists all saved CloudWatch Logs Insights queries
Tools
The server provides the following tools:
query_logs
- Query CloudWatch logs using CloudWatch InsightsParameters:
log_group_names
: Single log group name or list of log group names to queryquery_string
: CloudWatch Insights query stringstart_time
: (Optional) Start time for the query in Unix timestamp millisecondsend_time
: (Optional) End time for the query in Unix timestamp milliseconds
Features:
Automatically parses JSON in @message field
Returns structured data for JSON messages
Handles multiple log groups in a single query
discover_log_fields
- Discover available fields across multiple log groupsParameters:
log_group_names
: Single log group name or list of log group names to analyze
Features:
Efficiently discovers fields across multiple log groups
Assumes shared schema across log groups
Detects nested JSON fields in @message
Identifies field types (number, boolean, string, array)
log_group_exists
- Check if CloudWatch log groups existParameters:
log_group_names
: Single log group name or list of log group names to check
Returns:
Dictionary mapping each log group to its existence status
get_saved_queries
- Fetch all saved CloudWatch Logs Insights queriesNo parameters required
Testing the CloudWatch integration
You can test the CloudWatch integration directly using the provided test script:
Examples with MCP CLI
Using the MCP CLI:
License
MIT
This server cannot be installed
remote-capable server
The server can be hosted and run remotely because it primarily relies on remote services or has no dependency on the local environment.
A simplified MCP server that provides a streamlined way to interact with AWS CloudWatch resources (log groups, log queries, and alarms) through the MCP protocol.
Related MCP Servers
- AsecurityFlicenseAqualityAn MCP server that provides tools for interacting with AWS S3 buckets, enabling direct access to S3 operations through the Model Context Protocol.Last updated -1
- -securityFlicense-qualityAn MCP server that allows AI assistants to access AWS CloudWatch logs by listing log groups and reading log entries.Last updated -25
Log Analyzer with MCPofficial
-securityAlicense-qualityA Model Context Protocol server that provides AI assistants access to AWS CloudWatch Logs, enabling browsing, searching, summarizing, and correlating logs across multiple AWS services.Last updated -127Apache 2.0- -securityFlicense-qualityAn MCP server that enables interaction with Google Cloud Logging API, allowing users to write, read, and manage log entries and configurations through natural language.Last updated -