VocalHyre
Sign up

NoSQL · Data models beyond tables

How does Redis eviction work?

EasyredisevictioncachingUpdated Oct 2026
Quiz this track

Short answer

Redis enforces maxmemory with a policy such as noeviction, allkeys-lru, volatile-lru, allkeys-lfu or volatile-ttl. The policy decides which keys are removed when memory is full.

When memory reaches maxmemory, Redis applies the configured policy to free space. If no key can be evicted, it returns an error for writes. The choice between allkeys and volatile determines whether keys without a TTL are candidates.

text
maxmemory 4gb
maxmemory-policy allkeys-lru

maxmemory-policy noeviction
maxmemory-policy volatile-lru
maxmemory-policy allkeys-lfu
maxmemory-policy volatile-ttl
A cache usually uses allkeys-lru or allkeys-lfu. A store that also caches should use a volatile policy and only cache keys with a TTL.

Choosing a policy

  • noeviction: reject writes when full, safest when every key must survive
  • allkeys-lru: evict least recently used keys across the whole keyspace
  • allkeys-lfu: evict least frequently used keys, better for stable hot sets
  • volatile-lru or volatile-ttl: only evict keys with a TTL, protecting persistent data
  • random: cheap but unpredictable, rarely the right choice

Redis approximates LRU and LFU by sampling rather than maintaining a perfect ordering. That is why the policy is fast. It also means a recently used key can occasionally be evicted, so the cache must tolerate misses.

Do not use noeviction on a cache and let it run out of memory; writes will fail. Do not use allkeys-lru on a mixed store where some keys must never be evicted. Separate caches from durable data when their policies conflict.

Follow-ups you should expect

Interviewers use these to check whether the first answer was understanding or memorisation.

  1. 1What is the difference between LRU and LFU?
  2. 2Why does Redis approximate LRU?
  3. 3What happens with maxmemory-policy noeviction when memory is full?
  4. 4Why separate cache and durable keys?

Say it in the interview

Redis evicts keys when it reaches maxmemory, using a policy like allkeys-lru or volatile-ttl. For a pure cache I use allkeys-lru or lfu, because every key is disposable. For a mixed store I use a volatile policy and only give cache keys a TTL, so durable keys are protected. The eviction is approximate because Redis samples keys instead of keeping exact order, so the application must handle misses. I never run noeviction on a cache unless I want writes to fail at the limit.

Use this as a model, not a script. Saying it in your own words is what makes it survive follow-up pressure.

Related questions

EasyWhich Redis data structures would you use and when?MediumHow does Redis persistence work?HardHow do you control NoSQL costs in production?

Reading is step one

Now say it out loud.

Understanding an answer and delivering it under pressure are different skills. Rehearse this topic in a live NoSQL mock interview and get scored feedback on depth, clarity and confidence.

Start a free NoSQL interviewBrowse all questions