@quantabit/job-sdk (Job Queue SDK)

Universal message queue wrapper SDK supporting rapid migration from monolith to microservices. Internally uses a unified design pattern to seamlessly switch between Redis and RabbitMQ engines, implementing reliable task delivery, consumption, and retry mechanisms.

Features

• Multi-engine support (Redis and RabbitMQ)
• Ready to use: Built on bullmq (Redis) / amqplib (RabbitMQ), simplifying complex configurations
• Standard Payload: Auto-generated taskId, write timestamp throttle
• Built-in retry/delay queue: Redis engine supports delay and retries parameters directly
• Full-stack compatible: Works with Node.js, Next.js server, Nest, and Express projects

Installation

cd <your-project>
npm install @quantabit/job-sdk

Quick Start

1. Initialize Engine

Specify engine to switch the backend implementation at any time — business code remains transparent.

import { JobQueue } from "@quantabit/job-sdk";

// Use Redis (default, suitable for common microservices/task distribution)
const jobQueue = new JobQueue({
  engine: "redis",
  config: {
    connection: { host: "127.0.0.1", port: 6379 }, // ioredis config
  },
});

// Use RabbitMQ (suitable for high-throughput, complex dead-letter queue scenarios)
const rabbitQueue = new JobQueue({
  engine: "rabbitmq",
  config: {
    url: "amqp://127.0.0.1", // RabbitMQ connection string
  },
});

2. Publish Task (Publisher)

await jobQueue.connect();

// Unified publish behavior with auto-generated taskId, source info, built-in retry and delay options
const jobId = await jobQueue.publishTask(
  "deploy-queue",
  {
    projectId: "plans",
    repository: "org/repo",
    commitId: "abcdef",
  },
  {
    source: "github-intel-deploy", // Mark source
    retries: 3, // Failed retry count (Redis engine)
    delay: 2000, // Delay 2s before processing (requires backend support)
  },
);

console.log("Task assigned, JobID:", jobId);

3. Consume Task (Consumer)

Regardless of whether you use Redis or RabbitMQ, the handler function signature is consistent.

await jobQueue.connect();

jobQueue.consumeTask(
  "deploy-queue",
  async (data, context) => {
    // data : published payload { projectId, repository, commitId }
    // context: { taskId, source, timestamp, raw } - SDK provides standard context

    console.log(
      `[${context.source}] Received deploy task, Project: ${data.projectId}, TaskID: ${context.taskId}`,
    );

    // Simulate task processing logic; if an error is thrown, auto-enters failed retry pool
    // throw new Error("Deploy failed");
  },
  {
    concurrency: 5, // Concurrent processing count
  },
);

Best Practices & Use Cases

1. Unified Admin Backend: Dispatch async tasks directly via SDK (e.g., "rebuild full-text index", "batch send emails", "export Excel").
2. Eco Apps Deployment Pipeline: Listen and execute long-running tasks (e.g., running Python verify_deployment.py) and update cloud services.
3. Order System Deferred Processing: Use the delay option to auto-cancel unpaid orders after timeout.

🌐 Brand & Links

• Official Mainnet: QuantaBit Chain
• Developer Platform: Developer Platform
• Open Platform: Open Platform
• Payment Platform: Pay Platform
• Feedback: Feedback