All 7 Caching Patterns questions
More in Caching Patterns
What is the write through caching pattern, and how does it differ from cache aside? What is the write behind, also called write back, caching pattern, and what risk does it introduce? How does the read through caching pattern work, and how is it different from cache aside from the application's perspective? How do you decide which caching pattern, cache aside, write through, or write behind, is right for a specific application feature? How do you implement cache invalidation correctly when the same underlying data is cached under multiple different keys?
Advertisement
Back to results for “Distributed Locks with Redis”
Redis › Caching Patterns

How would you prevent a cache stampede, where many requests simultaneously try to regenerate the same expired cache entry at once?

Advanced

Answer

You can use a distributed lock so only one request regenerates the expired value while others wait briefly or serve a slightly stale version, or use a technique like probabilistic early expiration, where a small percentage of requests refresh the cache slightly before it actually expires, spreading out the regeneration load over time instead of all at once.
if redis.set(f'lock:{key}', 1, nx=True, ex=10):
    value = database.get_data()
    redis.set(key, value, ex=3600)
    redis.delete(f'lock:{key}')
else:
    -- wait briefly and retry, or serve stale data
    pass
Real-world example A popular news website prevents a cache stampede on its most viewed article by using a lock so only one request regenerates the cached article content when it expires, while other simultaneous visitors briefly see the slightly older cached version.

Common follow-ups: What is the difference between a cache stampede and normal cache misses?;How does probabilistic early expiration actually work in practice?

Distributed Locks with Redis;Redis Performance Tuning & Benchmarking
All 7 Caching Patterns questions