Distributed Task Scheduler .

Problem

Running background tasks reliably in a distributed environment without duplicate execution or missed jobs.

Solution

A task scheduler inspired by Celery and Sidekiq, with distributed locking to ensure only one worker picks up each task.

import redis
from contextlib import contextmanager

@contextmanager
def distributed_lock(lock_name, timeout=10):
    lock = redis_client.lock(lock_name, timeout=timeout)
    acquired = lock.acquire(blocking=False)
    
    try:
        yield acquired
    finally:
        if acquired:
            lock.release()

def execute_task(task_id):
    with distributed_lock(f"task:{task_id}") as acquired:
        if not acquired:
            return  # Another worker is handling this
        
        # Execute task
        process_task(task_id)

Features

• Cron Expressions: Standard cron syntax for scheduling
• Retries: Exponential backoff with max attempts
• Priorities: High/medium/low priority queues
• Dead Letter Queue: Failed tasks for manual review

Reliability

• Exactly-once execution guarantee
• Persistent queue in PostgreSQL
• Worker health checks and automatic restart