codecontext incident
Analyze code changes around production incidents
The incident
command helps you quickly identify code changes that might have caused production issues. It uses natural language time parsing and intelligent change analysis to pinpoint potential culprits.
Natural Time Parsing
"since 2 hours ago", "last Friday", "after 3pm yesterday"
Smart Analysis
Ranks changes by likelihood of causing the issue
Root Cause Fast
Find problematic changes in seconds, not hours
$ codecontext incident "login failures since 2 hours ago"
Interactive Demo
Natural Language Time Parsing
CodeContext understands natural language time expressions. Here are some examples:
Relative Time
"since 2 hours ago"
"in the last 30 minutes"
"after yesterday 3pm"
"since last Friday"
Specific Dates
"after January 15th 2pm"
"between 9am and 11am today"
"since Monday morning"
"during the weekend"
Options
--files, -f
Filter to specific file patterns
$ codecontext incident "errors today" --files "*.js"
--author, -a
Filter by commit author
$ codecontext incident "since 3pm" --author "john"
--limit, -l
Limit number of changes shown (default: 10)
$ codecontext incident "last hour" --limit 5
How It Works
Parse Time Expression
Converts natural language like "since 2 hours ago" to timestamps
Analyze Git History
Finds all code changes in the specified time range
Rank by Impact
Uses AI to assess likelihood each change caused the issue
Provide Context
Shows what changed, who changed it, and why it might be the culprit
Common Incident Queries
Performance Degradation
codecontext incident "slow API responses since this morning"
Feature Broken
codecontext incident "payment processing errors after 2pm deploy"
Integration Issues
codecontext incident "webhook failures in the last hour" --files "src/webhooks/*"
Database Problems
codecontext incident "connection timeout errors since yesterday" --files "*.sql,**/db/*"
Pro Features
Advanced Incident Analysis
Pro users get enhanced incident analysis capabilities:
- • Cross-repository incident correlation
- • Deployment tracking integration
- • Team notification when likely cause found
- • Historical incident pattern detection
- • Integration with monitoring tools
Best Practices
Be specific about timing
The more precise your time range, the better the results. "After 3:15pm" is better than "this afternoon".
Include error context
Mention specific errors or symptoms: "login 401 errors" helps identify auth-related changes.
Check deployment times
Correlate incident timing with deployment schedules for faster root cause identification.
See Also
- codecontext health - Monitor documentation health over time
- codecontext review - Review changes before they cause incidents
- Best Practices Guide - Best practices for using CodeContext during incidents