CloudWatch Application Signals

CloudWatch Application Signals

Key Tools

• audit_services - Primary service health auditing (use auditors="all" for root cause)
• audit_slos - SLO compliance monitoring
• audit_service_operations - Operation-specific analysis
• list_monitored_services / get_service_detail - Service discovery
• search_transaction_spans - 100% sampled OpenTelemetry spans
• query_sampled_traces - X-Ray traces (5% sampled)
• analyze_canary_failures - Synthetics canary analysis
• get_enablement_guide - Setup guidance

Best Practices

• Start with audit tools for overview, then drill down
• Use wildcard patterns for discovery: *payment*
• Use Transaction Search for 100% trace visibility (not X-Ray's 5%)
• Use auditors="all" for root cause analysis
• Always specify time ranges

Workflow: Service Health Investigation

1. audit_services with * to find issues
2. Deep dive into problematic service with auditors="all"

MCP Server Config

{
  "mcpServers": {
    "cloudwatch-appsignals": {
      "command": "uvx",
      "args": ["awslabs.cloudwatch-applicationsignals-mcp-server@latest"],
      "env": {
        "AWS_PROFILE": "default",
        "AWS_REGION": "us-east-1",
        "FASTMCP_LOG_LEVEL": "ERROR"
      }
    }
  }
}

Requires AWS credentials and Application Signals enabled in your account.