BullMQ

0 articles

JoyMini API — Enterprise-Level NestJS Backend Architecture Practice

A unified backend platform built with NestJS + Prisma + PostgreSQL + Redis + BullMQ, supporting 200+ API endpoints across 4 frontend applications

Project Cases

Translation Parallelism Optimization: Performance Leap from Serial to Concurrent Processing

This article analyzes the fundamental performance bottleneck of the blog translation system—serial processing caused by Worker concurrency=1. Through optimizations such as increasing concurrency and reducing internal latency, translation throughput is improved by approximately 5 times.

Interface Management

AI Translation Task Monitoring Dashboard: 1309 Lines Complete Engineering Implementation

Build an enterprise-level AI translation monitoring center—from BullMQ queue real-time monitoring, multi-dimensional progress tracking, to task history management and error alerts

Frontend Development

QueueMonitorService: BullMQ Queue Monitoring Dashboard

The JoyMini API queue monitoring service builds a real-time dashboard based on the BullMQ event system, tracking queue status, Job progress, failure rates, and processing latency, supporting visual operations in the admin backend.

Interface Management

Media Processing Pipeline: Sharp Image Compression + FFmpeg HLS Video Transcoding

The JoyMini API's media processing pipeline uses Sharp for multi-format image variant generation and BlurHash encoding, and FFmpeg for HLS adaptive video transcoding, all asynchronously driven by BullMQ.

Interface Management

GroupService Group Buying System: Redis Optimistic Lock + BullMQ Settlement Queue

JoyMini's group buying system is based on Redis optimistic locking to prevent overselling and a BullMQ settlement queue for asynchronous profit distribution, covering the complete transactional flow of group creation, joining, automatic group formation, and settlement.

Interface Management

API File Upload System — Cloudflare R2 S3 + BullMQ Media Processing Pipeline

JoyMini API's file upload system is based on Cloudflare R2 S3 object storage and BullMQ asynchronous queues, covering a complete media processing pipeline including image compression (Sharp), video HLS transcoding (FFmpeg), and BlurHash generation.

Interface Management

API Background Task System — BullMQ Queue Architecture and Worker Implementation

JoyMini API's background task system built on BullMQ (Redis-driven queue), covering the architecture design of 3 queues and 4 Workers, including the complete implementation of image compression pipeline, Sharp composition, video HLS transcoding, and thumbnail extraction.

Interface Management

Avatar Service, Payment Integration, and Public Cache Interceptor

JoyMini API's automatic avatar generation service (Sharp compositing + BullMQ queue), Xendit payment integration (with distributed lock reconciliation), and a Redis-based public cache interceptor implementation.

Interface Management