For Problem Solvers

Debug

Systematic Debugging

Binary search, divide and conquer, cause elimination—structured methodologies to find bugs.

The Transformation

Before scenario
(Coming soon)

After transformation
(Coming soon)

See It In Action

Scenario: “Why is user engagement dropping after the latest release?”

Issue Under Investigation

Users report 500 errors on checkout, but only on Tuesdays between 2-4pm EST

Approach: Binary Search

Systematic Investigation

1
Investigation

Check if error occurs in payment service OR order service

Observation

Logs show payment service returns 200, order service throws exception

Scope Narrowed
Remaining: Order service (3 modules)
2
Investigation

Isolate: inventory module vs pricing module vs fulfillment module

Observation

Pricing module queries external tax API that times out

Scope Narrowed
Remaining: Tax API integration
3
Investigation

Test tax API directly during failure window

Observation

Tax API has scheduled maintenance Tue 2-4pm - returns 503

Root Cause Found
Remaining: External dependency
Root Cause Identified

External tax API weekly maintenance window not accounted for in integration design.

Resolution

Add circuit breaker with cached tax rates fallback. Alert on tax API degradation.

Binary Search Debugging • 3 Steps to Root Cause

When to Use Debug

  • 1
    Investigating production issues methodically
  • 2
    Finding root causes of complex bugs
  • 3
    When trial-and-error isn't working
  • 4
    Performance bottleneck identification
  • 5
    Regression debugging after updates

Works Well With

Pro tip: Chain Debug with Trace → Hypothesis for comprehensive analysis.

Ready to use Debug?

Add Think to your AI assistant in one command

npx @anthropic-ai/claude-code mcp add think-mcp
View all tools•GitHub