인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
생산성 및 협업 · SKILL.md

nestjs-scheduling

nestjs-scheduling

Implement distributed cron jobs with Redis-based locking and BullMQ offloading in NestJS. Use when adding @Cron scheduled tasks, preventing duplicate runs across pods, or delegating heavy work to queue workers.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: nestjs-scheduling description: Implement distributed cron jobs with Redis-based locking and BullMQ offloading in NestJS. Use when adding @Cron scheduled tasks, preventing duplicate runs across pods, or delegating heavy work to queue workers. metadata: triggers: files: - '**/*.service.ts' keywords: - "@Cron" - CronExpression - ScheduleModule

Task Scheduling & Jobs

Priority: P1 (HIGH)

Workflow: Add Scheduled Task

  1. Register ScheduleModule — Import ScheduleModule.forRoot() in AppModule.
  2. Create cron handler — Decorate service method with @Cron(CronExpression.*).
  3. Add distributed lock — Apply Redis lock decorator to prevent multi-pod duplication.
  4. Offload heavy work — Push job IDs to BullMQ; let workers process them.
  5. Wrap in try/catch — Uncaught exceptions in cron handlers crash entire Node process.
  6. Verify — Test with 2+ instances to confirm only one acquires lock.

Problem & Solution

  • Problem: @Cron() runs on every instance. In K8s with 3 pods, your "Daily Report" runs 3 times.
  • Solution: Distributed Locking using Redis.
  • Pattern: Using decorator to wrap cron method.
  • Logic: SET resource_name my_random_value NX PX 30000 (Redis Atomic Set).

Cron Decorator Pattern

  • Implementation:

See implementation examples

  • Tools: Use nestjs-redlock or custom Redis wrapper via redlock library.

Cron-to-Queue Offload

See implementation examples

Job Robustness

  • Isolation: Never perform heavy processing inside Cron handler.
  • Pattern: Cron -> Push Job ID to Queue (BullMQ) -> Worker processes it.
  • Why: Cron schedulers can get blocked by Event Loop; Workers scalable.
  • Error Handling: Wrap ALL cron logic in try/catch. Uncaught exceptions in Cron job can crash entire Node process.

Anti-Patterns

  • No unguarded cron logic: Always wrap in try/catch; uncaught exceptions crash entire Node process.
  • No direct cron processing: Push to BullMQ queue; workers scalable, cron handlers not.
  • No bare @Cron in multi-pod: Use distributed locking (redlock) to prevent duplicate concurrent runs.

Canonical response anchors

When this skill applies, preserve the following domain terminology or equivalent concrete examples in the answer when relevant:

  • crash

Remediation anchors

  • Remediation anchors: BullMQ, event loop, worker
저장한 Skills