feat(phase4-batch4.1): Centralized error handling & DLQ

[arpayidcloud-byte]

Summary

Implement Phase 4.1 - Error Handling & Dead Letter Queue per MASTER_PLAN_PRODUCTION.md

Changes

• Centralized ErrorHandler with classification & recovery
• Error severity levels (CRITICAL/HIGH/MEDIUM/LOW)
• Error categories (TRANSIENT/CONFIGURATION/VALIDATION/EXTERNAL/INTERNAL/RESOURCE_EXHAUSTED)
• Recovery strategies (RETRY_EXPONENTIAL/RETRY_FIXED/DEAD_LETTER/FAIL_IMMEDIATE/CIRCUIT_BREAKER)
• Dead Letter Queue CLI commands (list/clear/size)
• Integration with existing CircuitBreaker
• Error boundary wrapper for safe execution

Error Classification

Category	Retryable	Strategy	Example
TRANSIENT	✅	RETRY_EXPONENTIAL	Timeout, rate limit, network
CONFIGURATION	❌	FAIL_IMMEDIATE	Missing API key, invalid config
VALIDATION	❌	FAIL_IMMEDIATE	Invalid arguments
EXTERNAL	✅	RETRY_EXPONENTIAL	Provider errors
RESOURCE_EXHAUSTED	✅	CIRCUIT_BREAKER	Quota exceeded
INTERNAL	❌	DEAD_LETTER	Unknown errors

CLI Commands

# List failed tasks
pnpm agentx dlq list

# Show DLQ size
pnpm agentx dlq size

# Clear DLQ
pnpm agentx dlq clear

CI Status

• Typecheck ✅
• Lint ✅
• Build ✅
• Tests ⏳
• Coverage ⏳

Files Changed

• packages/runtime/runtime-production/src/error-handler.ts (NEW - 395 lines)
• apps/cli/src/commands/dlq.ts (NEW - 91 lines)
• apps/cli/src/index.ts (MODIFIED - register dlq command)
• README.md (MODIFIED - add DLQ commands)

Related

• Phase 3.3: PR feat(phase3-batch3.3): E2E demo implementation #61 (E2E Flow) ✅ Merged
• Phase 4.1: This PR (Error Handling & DLQ) ⬅️ CURRENT
• Phase 4.2: Circuit Breakers (already implemented)
• Phase 4.3: Retry Logic (already implemented)
• Phase 4.4: Graceful Shutdown (next)

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.