- HealthWatcher thread: monitors proxy /health every 5s - LogAnalyzer thread: tails cc-debug.log for 18 failure signal patterns - Tier 1 rule engine: 14 rules for instant auto-recovery (< 1s) - Tier 2 incident store: JSON pattern database with success rates - Tier 3 AI diagnostic agent: calls configurable provider/model for novel failures - AIMonitoringWindow GUI: ON/OFF toggle, provider/model/API key selector, incident log - 30 fault types catalogued across 5 categories (A-E) - Enhanced /health endpoint with memory_mb, uptime_s, requests_total - Auto-restart proxy, auto-clear schema cache, kill stale processes - Safety: rate-limited AI calls, restart caps, cooldowns per pattern - AI Monitoring design spec (AI-MONITORING-DESIGN.md) - 54 self-test patterns passing
1.6 KiB
Executable File
1.6 KiB
Executable File