Summary of Redis Hot Key and Big Key Problem

1. What is Redis hot key?

Hot key (Hot Key) definition:

The key that is frequently accessed (read/write) within a unit time will lead to concentrated access and excessive pressure.

Hot key Common manifestations:

• Extremely high QPS (a key is accessed tens of thousands of times per second)
• A business reads and writes a key at high frequency (such as flash sale inventory, ranking top1)
• Redis instance CPU exception,But only serve one key

Risks of hot keys:

question	illustrate
Excessive pressure on a single point	All requests are called to the same key
Cache breakdown	Hot key expires, a large number of requests fall to the DB
Master-slave replication delay	Hot key changes frequently → Master-slave synchronous data increases sharply
Node imbalance (cluster)	The key distribution in Redis Cluster is uneven

Hot key solution:

question	illustrate
Excessive pressure on a single point	All requests are called to the same key
Cache breakdown	Hot key expires, a large number of requests fall to the DB
Master-slave replication delay	Hot key changes frequently → Master-slave synchronous data increases sharply
Node imbalance (cluster)	The key distribution in Redis Cluster is uneven

2. What is Redis big key?

Big key (Big Key) definition:

Refers tovalue too big(For example, a Hash contains millions of fields, or a List has more than 100,000 items) or a ** string with a large size (such as a few MB images)** key.

Risks of big keys:

question	illustrate
❗ Delete blocking	DELLarge key blocks Redis single thread
❗ Master-slave copying slows down / lost	The master node transmits a large key → replication delay
❗ Causes RDB/AOF to surge	Dump will be stuck once
❗ Affects the performance of a single command	The operation of a large key will slow down, for exampleHGETALL

Big key recognition method:

• useredis-cli --bigkeysCommand scan instance

• useredis-rdb-toolsAnalyze RDB files

• View via slow query log + monitoringHGETALL、LRANGE、SMEMBERSThe big key

Big key solution:

plan	illustrate
✅ Split data structure	For example, a large hash is split into multiple small hashs (click ID)
✅ Control the maximum number of fields/elements	Controlling individual structure members ≤ 10K
✅ Disable Dangerous Commands	For example, closeKEYS, FLUSHALL, HGETALL
✅ Delayed/slow delete strategy	For example, large keys are divided into batchesUNLINKdelete
✅ Limit maximum value size	Do not exceed a few KB of strings, avoid exceeding MTU

The difference between UNLINK vs DEL:

Order	illustrate
DEL	Delete immediately (blocking thread)
UNLINK	Asynchronous deletion,Non-blocking ✅

✅ Redis 4.0+ It is recommended to use UNLINK to delete large keys!

Summary comparison table

type	Trigger method	risk	Solution core
Hot key	Access the same key at high frequency	CPU soaring, breakdown, hot spots	Local cache, sharding, current limiting, preheating
Big key	value structure too large	Slow query, blocking, synchronization issues	Segment splitting, asynchronous deletion, structural constraints, UNLINK

This is all about this article about Redis hot key and big key issues. For more related Redis hot key and big key content, please search for my previous articles or continue browsing the related articles below. I hope everyone will support me in the future!