Messaging Infrastructure

Queue Management with BullMQ

Overview

We use BullMQ for job queuing and processing, providing:

• Horizontal scaling capabilities
• Message replay and retry mechanisms
• Distributed job processing
• Redis as the backend storage

Queue Monitoring

BullBoard URLs

• Staging: https://bull-board-staging-abda.up.railway.app/
• Production: https://creative-warmth-production.up.railway.app/

Important Notes

• Adding a new queue requires a BullBoard service restart
• Jobs are created and processed using BullMQ
• Limitation:
  • Consumer groups feature is NOT available (paid feature)
  • 1 will be consumed once, meaning we CAN'T currently gave 2 separate process consuming the same message for two different tasks

Job Processing

Retry Mechanisms

• Retry configuration can be customized per processor
• Default retry strategies can be configured
• Supports backoff and custom error handling

Best Practices

• Design idempotent job processors
• Use appropriate retry and backoff strategies
• Monitor queue performance and job statuses
• Handle potential race conditions and duplicates
• Create types for Jobs inputs and outputs

Infrastructure Details

• Backend: Redis
• Queue Management: BullMQ
• Monitoring: BullBoard

Future Improvements

• Explore advanced consumer group strategies
• Enhance job monitoring and alerting
• Investigate more sophisticated retry mechanisms