feat(resilience): implement circuit breakers, bulkheading, and failover for external services

[dDevAhmed]

Summary

Implements robust circuit breakers, bulkheading patterns, graceful degradation, failover mechanisms, and automated recovery procedures across all external service calls to prevent cascading failures and ensure system stability.

What Changed

• Registered CircuitBreakerRecoveryService in ErrorHandlingModule to enable automated recovery of open circuit breakers via cron-based health checks
• Added circuit breaker + bulkhead to EmailService (SMTP/nodemailer) with graceful degradation — returns false instead of throwing when email delivery fails
• Added circuit breaker + bulkhead to SmsService (Twilio) with graceful degradation — returns false instead of throwing when SMS delivery fails, and degrades when Twilio client is not initialized
• Added failover mechanism to NotificationProcessor — when email or SMS fails, automatically attempts push notification delivery as a fallback channel
• Updated NotificationsModule to provide CircuitBreakerService, BulkheadService, and CorrelationIdService
• Completed circuit breaker coverage for blockchain services (Ethereum RPC, Stellar Horizon) with bulkhead isolation
• Completed circuit breaker coverage for FCM mobile push with graceful degradation
• Added unit tests for CircuitBreakerService (register, execute, getState, getMetrics, reset)
• Added unit tests for BulkheadService (createBulkhead, execute, getMetrics, resetMetrics, removeBulkhead)
• Added chaos engineering tests verifying cascading failure prevention, component isolation, graceful degradation, automated recovery, and failover under failure conditions

Why

Without circuit breakers and bulkheads, failures in external services (SMTP, Twilio, Ethereum RPC, Stellar Horizon, FCM) cascade through the system, causing widespread outages. This implementation ensures:

• Failing components are isolated from the rest of the system
• The system degrades gracefully (returns fallback values) rather than failing completely
• Notifications failover to alternative channels when primary channels are unavailable
• Circuit breakers automatically recover after a cooldown period

Testing Performed

• Unit tests for CircuitBreakerService (8 test cases)
• Unit tests for BulkheadService (9 test cases)
• Chaos engineering resilience tests (12 test cases across 6 scenarios)
• Lint
• Build

Edge Cases Considered

• Circuit breaker opens after threshold failures and prevents cascading calls to failing services
• Bulkhead limits concurrent requests to prevent resource exhaustion
• Email/SMS services return false (not throw) when circuit is open — enabling graceful degradation
• Notification failover records the fallback channel in notification metadata
• Twilio client not initialized returns false instead of throwing
• Automated recovery service monitors and resets open circuit breakers after cooldown

Risks

None. All changes are additive — existing API contracts are preserved. Circuit breakers and bulkheads are internal infrastructure that do not change external API behavior.

Closes #403