Mastering Redis Hotspot Keys: Detection, Risks, and Solutions

Preface

A reader preparing for a Meituan interview was asked how to solve Redis hotspot keys; the article outlines a comprehensive answer.

1. What Is a Hotspot Key

In Redis, a key that receives a very high access frequency is called a hotspot key . Excessive requests to a hotspot key can exhaust server resources and even cause crashes.

2. Problems Caused by Hotspot Keys

CPU overload : Frequent access can overload the Redis instance’s CPU, degrading overall performance.

Memory pressure : Large hotspot keys consume significant memory, potentially leading to memory exhaustion.

Load imbalance : Hotspot keys may concentrate on one or few cluster nodes, causing uneven load.

Master‑slave sync delay : High load on the master can increase replication lag, risking data inconsistency.

Cache breakdown : When a hotspot key expires, a flood of requests may bypass the cache, overwhelming downstream services.

3. Common Causes of Hotspot Keys

Popular data : Frequently accessed items such as hot product pages or trending posts.

Short‑term spikes : News bursts, flash sales, etc., cause sudden traffic spikes.

Request sharding concentration : Fixed key names may hash to the same node in a cluster, creating a hotspot.

4. How to Monitor Hotspot Keys

MONITOR command : Streams all commands in real time; useful for debugging but has high overhead.

Custom statistics scripts : Use Lua or combine commands like

SCAN

and

TTL

to periodically record high‑frequency keys.

Redis built‑in analysis : From Redis 4.0,

redis-cli --hotkeys

helps identify hotspot keys.

Third‑party monitoring tools : RedisInsight, Prometheus, etc., can alert on abnormal key access patterns.

SLOWLOG command : Records slow commands; frequent operations on a key may surface here.

5. How to Identify Hotspot Keys

Experience‑based judgment of likely hot data (e.g., product detail pages).

Client‑side statistics reporting via Redis clients (Jedis, Lettuce).

Proxy‑layer reporting when using Redis proxies such as Twemproxy, Codis, or Redis Cluster.

6. How to Solve Hotspot Keys

Common approaches include expanding the Redis cluster, distributing hotspot keys across servers, and using a second‑level cache.

6.1 Redis Cluster Expansion

Adding shards and replicas spreads load, especially read traffic, improving concurrency.

Advantages : Horizontal scaling handles higher load.

6.2 Distribute Hotspot Keys

Modify key patterns (e.g., add random prefixes) to split a hot key into multiple keys, balancing traffic across nodes.

Advantages : Avoids single‑node bottlenecks and increases overall throughput.

6.3 Use a Second‑Level Cache

Introduce local caches (Guava, Caffeine) in the application layer to store hot data in JVM memory, reducing Redis reads.

Advantages : Lowers Redis pressure and improves response time.

7. Backend Thinking Inspired by Hotspot Keys

The discussion leads to broader backend concepts such as cache design and load‑balancing strategies.

7.1 Cache Design

Adopt multi‑level caching (local + Redis + CDN) with appropriate expiration policies to protect backend databases.

Consider a multi‑tier cache architecture to reduce database load while ensuring data consistency.

7.2 Load Balancing and Sharding Strategies

Implement effective load‑balancing across services and databases; optimize sharding algorithms to distribute data evenly and prevent hotspots.