Objective

Create an automated workflow that investigates CI/CD failures, analyzes logs, identifies root causes, and creates detailed investigation reports with remediation steps.

Context

This repository has complex Docker/networking tests that frequently fail with opaque errors like "subnet pool overlap" and container cleanup race conditions. Manual log analysis wastes developer time and accumulated failure patterns aren't leveraged.

Approach

1. Create workflow file: .github/workflows/ci-doctor.md
2. Configure triggers:
   • workflow_run on completion of: test-integration, test-coverage, test-action
   • Only trigger when conclusion == 'failure'
3. Implement investigation protocol:
   • Fetch workflow run details and job logs via GitHub API
   • Analyze for Docker network issues (subnet pool exhaustion, overlaps)
   • Check for container cleanup race conditions
   • Detect iptables rule conflicts
   • Identify Squid proxy startup failures
   • Search for similar past failures using cache-memory
4. Create investigation report:
   • Detailed root cause analysis
   • Specific remediation steps
   • Link to similar past failures
   • Label with bug, ci, needs-investigation
5. Store failure patterns: Update cache-memory with new patterns

Files to Create

• .github/workflows/ci-doctor.md - Main workflow
• Reference: agentics/ci-doctor.md template

Domain-Specific Focus Areas

• Docker network pool exhaustion (172.30.0.0/24 subnet conflicts)
• Container cleanup race conditions (timeout kills leave orphaned resources)
• iptables rule conflicts (NET_ADMIN capability issues)
• Squid proxy healthcheck failures
• GitHub Actions runner Docker version incompatibilities

Acceptance Criteria

• Workflow triggers automatically on CI failures for specified test workflows
• Creates detailed investigation issues with root cause analysis
• Identifies Docker/networking specific failure patterns
• Searches cache-memory for similar historical failures
• Provides actionable remediation steps
• Completes within 10 minutes timeout

Success Metrics

• Reduce failure diagnosis time from manual (hours) to <30 minutes
• Build knowledge base of 10+ common failure patterns within first month
  Related to [plan] Enhance agentic workflow maturity to Level 4 (Optimized) #332

AI generated by Plan Command for discussion #328