What You Can Do by Querying the Application Signals MCP

AIOps enhancements continue to roll out in CloudWatch Application Signals, and expectations are running high.

To see what the Application Signals MCP can actually do, I sent the following query exactly as described in the AWS documentation.

What questions can I ask about CloudWatch or Application Signals MCP Servers?

I received the response below, so I’m keeping it here as a personal reference.

---

⏺ Great! The MCP server for AWS Application Signals is now available.

You can ask questions like the following:

🎯 Service Audit / Health Check

Basic service audits

• Audit the health of all services
• Check the status of payment-related services
• Audit all services running on the EKS cluster
• Check the service status from last night

Detailed analysis by operation

• Audit the latency of the GET operation on the payment service
• Investigate the performance of a specific API endpoint
• Identify operations with high error rates

📊 SLO (Service Level Objective)

SLO management

• List all configured SLOs
• Check for SLO violations
• Show me the detailed configuration of a specific SLO
• Perform a root cause analysis of payment-related SLO violations

🔍 Trace / Log Analysis

Problem investigation

• Investigate the cause of errors occurring in a specific service
• Analyze the cause of high latency using traces
• Identify dependency issues
• Find outlier hosts

Log search

• Search error logs for a specific time window
• Investigate communication errors between services

📈 Metrics / Performance Analysis

Service metrics

• Retrieve the latency metrics of a specific service
• Check the trend of error rates over time
• Analyze performance trends over the past week

🚨 Failure Analysis / Root Cause Investigation

Incident response

• Investigate new errors that appeared after a deployment
• Identify issues with inter-service dependencies
• Audit dependencies on S3
• Check the quota usage of Tier 1 services

🕸️ Synthetics Canary Analysis

Canary monitoring

• Analyze canary failures for [canary name]
• Investigate canary failure patterns

Concrete Example Questions

• Overall service health check
  “Audit the health of all services”

• Detailed analysis of a specific service
  “Perform a root cause analysis of the latency issue on the GET operation of payment-service”

• Investigating SLO violations
  “Investigate SLO violations since yesterday and analyze the root cause”

• Identifying performance problems
  “Identify services with slow response times and check their error logs”

• Investigating dependency issues
  “Identify services that have connectivity problems with S3”

That’s all.

I hope you find this helpful.