agentcore-investigation

Correlate CloudWatch Logs Insights and OpenTelemetry spans to build AWS Bedrock AgentCore session timelines.

9.6k|1.7k|Updated Mar 21, 2025
One-click install
npx skills add https://github.com/awslabs/mcp --skill agentcore-investigation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agentcore-investigation
Source: https://github.com/awslabs/mcp/tree/main/src/cloudwatch-mcp-server/skills/agentcore-investigation
Command: npx skills add https://github.com/awslabs/mcp --skill agentcore-investigation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps you quickly diagnose and resolve issues within AWS Bedrock AgentCore runtime sessions by analyzing CloudWatch Logs Insights and OpenTelemetry spans.

Core Features & Use Cases

  • Session Debugging: Pinpoint errors, trace tool calls, and analyze latency in AgentCore sessions.
  • Log Analysis: Query CloudWatch Logs Insights, filter out noisy OpenTelemetry data, and build a clear timeline of events.
  • Use Case: When an AgentCore session fails or behaves unexpectedly, use this Skill to trace the execution flow, identify the exact point of failure, and understand the root cause by examining logs and spans.

Quick Start

Investigate session abcdef12-3456-7890-abcd-ef1234567890 using CloudWatch Logs Insights.

Frequently Asked Questions about agentcore-investigation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I debug AWS Bedrock AgentCore sessions using CloudWatch Logs?

Debug AgentCore sessions by querying CloudWatch Logs Insights and correlating OpenTelemetry spans. This filters noisy data, analyzes token usage and errors, and builds a timeline to identify the exact point of failure.

What is the best way to trace tool invocations and token usage in Bedrock AgentCore?

Trace tool invocations and token usage by correlating OpenTelemetry spans with CloudWatch Logs Insights. This filters noisy data and resolves session IDs to trace IDs for a clear execution timeline.

Do I need OpenTelemetry data to investigate AgentCore runtime session failures?

Yes, investigating AgentCore session failures requires both CloudWatch Logs Insights and OpenTelemetry data. Correlating these sources resolves session IDs to trace IDs and builds a session timeline.

Why does my AgentCore session fail or behave unexpectedly and how do I find the root cause?

Find the root cause of unexpected behavior by querying CloudWatch Logs Insights and correlating OpenTelemetry spans. This traces the execution flow, filters noise, and identifies the exact point of failure.

Can I analyze latency and filter noisy OpenTelemetry data in CloudWatch for AgentCore?

Yes, you can filter noisy OpenTelemetry data and analyze latency by querying CloudWatch Logs Insights. Correlating spans resolves trace IDs and builds a clear timeline of session events.

What limitations exist when debugging AgentCore sessions with CloudWatch Logs Insights?

The main limitation is the strict dependency on both CloudWatch Logs Insights and OpenTelemetry data. Without both sources, the Skill cannot resolve session IDs to trace IDs or build a timeline.