---
title: "How to prevent Cache Stampede (Thundering Herd) in Redis using Distributed Locks?"  
description: "How to prevent Cache Stampede (Thundering Herd) in Redis using Distributed Locks?"  
author: "Ravi Vishwakarma"  
published: 2026-09-17  
canonical: https://answers.mindstick.com/qa/117198/how-to-prevent-cache-stampede-thundering-herd-in-redis-using-distributed-locks  
category: "Redis"  
tags: ["Redis", "Caching", "Python", "performance"]  
reading_time: 1 minute  

---

# How to prevent Cache Stampede (Thundering Herd) in Redis using Distributed Locks?

A **cache stampede** (also known as the thundering herd problem) occurs when multiple concurrent client requests experience a cache miss at the exact same time. This causes all requests to hit the backend database simultaneously, leading to sudden CPU spikes and database exhaustion.

## Solution using Distributed Locks

By implementing a **distributed lock** alongside a standard **Cache-Aside pattern**, only the first thread that experiences a cache miss is allowed to query the database and rebuild the cached value, while subsequent concurrent requests wait or retry.

```python
import redis
import time

# Initialize Redis client connection
r = redis.Redis(host='localhost', port=6379, db=0)

def get_cached_data(key):
    # Try fetching data directly from cache
    val = r.get(key)
    if val:
        return val.decode('utf-8')

    # Try acquiring distributed lock with 10 sec timeout
    lock_acquired = r.set('lock:' + key, 'true', nx=True, ex=10)
    if lock_acquired:
        try:
            # Fetch fresh data from persistent database
            data = "computed_db_result"
            # Cache fresh value with 60 second TTL
            r.setex(key, 60, data)
            return data
        finally:
            # Release distributed lock safely
            r.delete('lock:' + key)
    else:
        # Sleep briefly and retry cache lookup
        time.sleep(0.1)
        return get_cached_data(key)
```


---

Original Source: https://answers.mindstick.com/qa/117198/how-to-prevent-cache-stampede-thundering-herd-in-redis-using-distributed-locks

Copyright © MindStick Software Pvt. Ltd. This Markdown version is provided for developers, AI systems, and offline reading.