AWS – Amazon CloudWatch and Application Signals MCP servers for AI-assisted troubleshooting

Today, AWS announces two new Model Context Protocol (MCP) servers in the AWS Labs MCP open-source repository: CloudWatch MCP server and Application Signals MCP server. These servers enable AI agents to leverage comprehensive observability capabilities for automated troubleshooting and monitoring. The MCP servers allow AI assistants to analyze metrics, alarms, logs, traces, and service health data across your AWS environment to quickly identify and diagnose issues through simple conversational interfaces.

The MCP servers provide curated sets of tools designed specifically for operational troubleshooting scenarios. The CloudWatch MCP server supports alarm-based incident response, metric analysis, and log pattern detection, while the Application Signals MCP server enables service health monitoring through Service Level Objectives (SLOs), and automated root cause analysis using OpenTelemetry data. By leveraging the MCP standard, AI agents can perform complex troubleshooting workflows through natural language interactions, from analyzing alarm patterns, and metric anomalies to investigating service health issues and querying logs and traces. Rather than requiring developers to manually navigate multiple AWS consoles and APIs, these MCP servers enable AI agents to orchestrate these interactions intelligently while reducing the development times typically required for API integrations.

The CloudWatch MCP server can be used with CloudWatch in all AWS regions, and the Application Signals MCP server can be used in all regions where Application Signals is available.

To download and try out these open-source MCP servers locally with your AI-enabled IDE of choice, visit the AWS Labs MCP open-source repository.

Read More for the details.