fix: resolve assigned issues #599, #600, #785, #786

[JamesVictor-O]

Summary

[CONTRACT] #599 - Optimize Storage Footprint for Large Employee Maps

• Compressed BatchStatusMap: Packs payment statuses into 2-bit-per-status format (16 statuses per u32 word), reducing storage from O(N) individual keys to a single key per batch
• archive_batch_statuses(): Admin function to compress PaymentEntry records into a BatchStatusMap after batch processing is complete, removing individual temporary storage keys
• get_archived_status(): Read payment status from compressed storage by batch_id and payment_index
• reduce_batch_ttl(): Apply shorter TTL to old batch records for faster archival
• Tiered TTL constants: ARCHIVE_TTL_THRESHOLD (5,000) and ARCHIVE_TTL_EXTEND_TO (50,000) for historical data vs active data

[CONTRACT] #600 - Implement Automatic Distribution Account Re-funding

• set_refund_config(): Admin configures distribution account, funding source, token, threshold, and refund amount
• check_and_refund(): Checks distribution account balance against threshold and automatically transfers refund_amount from funding source when below threshold
• get_refund_config() / remove_refund_config(): Config management functions
• Events: AccountRefundedEvent and RefundConfigUpdatedEvent for on-chain audit trail
• New error codes: AutoRefundFailed(26), RefundThresholdInvalid(27), RefundConfigNotSet(28), RefundNotNeeded(29)

[BACKEND] #785 - Backend Legacy Issue - Maintenance & Stability

• asyncHandler utility: Wraps async Express handlers to forward rejected promises to the error middleware, eliminating uncaught promise rejections across all controllers
• HttpError class: Factory methods (badRequest, unauthorized, notFound, etc.) for standardized error throwing that integrates with the centralized error handler
• preflightCheckWithAutoRefund(): New BalanceService method that attempts on-chain auto-refund before failing a preflight check, integrating with Issue [CONTRACT] Implement Automatic Distribution Account Re-funding #600's contract functionality
• Retry on Stellar submissions: withRetry() wrapper (3 attempts, exponential backoff) around StellarService.submitTransaction() in the payroll worker
• Standardized timeout errors: Response timeout now uses apiErrorResponse format instead of ad-hoc JSON

[BACKEND] #786 - Backend Legacy Issue - Maintenance & Stability

• Graceful worker shutdown: New stopWorkers() function that cleanly closes all 4 BullMQ workers (payroll, notification, scheduler, tx-verification)
• Server shutdown integration: Workers are now stopped before database pool closure during SIGTERM/SIGINT
• Webhook retry interval cleanup: clearInterval() on shutdown prevents timer leaks
• Database migrations: 050_auto_refund_audit_log.sql and 051_batch_archive_tracking.sql for off-chain tracking

Test plan

• All 121 Soroban contract tests pass (cargo test -p bulk_payment) including 15 new tests
• All 21 new backend unit tests pass (asyncHandler, HttpError, BalanceService)
• No TypeScript compilation errors in modified files
• Existing test suite not affected (no regressions)

Closes #599
Closes #600
Closes #785
Closes #786

[drips-wave]

@JamesVictor-O Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits