@meridianjs/job-queue-redis

Production-grade cron scheduler for MeridianJS, backed by BullMQ + Redis. Jobs persist across restarts, run on a distributed schedule, and are retried automatically on failure.

For local development without Redis, use @meridianjs/job-queue-local instead.

Installation

npm install @meridianjs/job-queue-redis

Configuration

Register in meridian.config.ts:

export default defineConfig({
  modules: [
    {
      resolve: "@meridianjs/job-queue-redis",
      options: { url: process.env.REDIS_URL },
    },
  ],
})

| Option | Type | Default | Description | |---|---|---|---| | url | string | — | Redis connection URL (required) | | prefix | string | "meridian:job" | Prefix for all BullMQ queue names |

Writing Jobs

Jobs are identical in both the local and Redis implementations — no code changes needed when switching:

// src/jobs/daily-summary.ts
import type { JobConfig } from "@meridianjs/types"

export default async function handler({ container }: { container: any }) {
  const logger = container.resolve("logger") as any
  logger.info("Sending daily summary emails")
  // ... your logic here
}

export const config: JobConfig = {
  name: "daily-summary",
  schedule: "0 8 * * *",  // every day at 8 AM
}

Each job gets its own named BullMQ queue (meridian:job:<name>). On restart, existing repeatable jobs are cleaned up and re-registered to prevent duplicates.

Behaviour

• Each job runs with concurrency: 1 — jobs of the same name run sequentially.
• Completed jobs are retained for the last 100 entries; failed jobs for 50.
• Worker and queue connections are gracefully closed on SIGTERM / SIGINT.

Switching from Local

// development
{ resolve: "@meridianjs/job-queue-local" }

// production
{ resolve: "@meridianjs/job-queue-redis", options: { url: process.env.REDIS_URL } }

License

MIT