Task Scheduling & Jobs

Distributed Cron (Critical)

• Problem: @Cron() runs on every instance. In K8s with 3 pods, your "Daily Report" runs 3 times.
• Solution: Distributed Locking using Redis.

- Pattern: Using a decorator to wrap the cron method. - Logic: SET resourcename myrandom_value NX PX 30000 (Redis Atomic Set).

Cron Decorator Pattern

• Implementation:

``typescript @Cron(CronExpression.EVERYMINUTE) @DistributedLock({ key: 'sendemails', ttl: 5000 }) async handleCron() { // Only runs if lock acquired } ``

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

Job Robustness

• Isolation: Never perform heavy processing inside the Cron handler.

- Pattern: Cron -> Push Job ID to Queue (BullMQ) -> Worker processes it. - Why: Cron schedulers can get blocked by the Event Loop; Workers are scalable.

• Error Handling: Wrap ALL cron logic in try/catch. Uncaught exceptions in a Cron job can crash the entire Node process.