10 Real-World Redis Use Cases Every Backend Engineer Should Know

Preface

Redis is a high‑performance key/value cache widely used in both production and interview settings. This article shares ten real‑world Redis scenarios that can help you solve common problems efficiently.

1. Count Page Visits

Websites often need to count homepage visits. Storing a single counter in a database and aggregating later is cumbersome. Define a Redis key, e.g., OFFICIAL_INDEX_VISIT_COUNT, and use the INCR command to increment it. For larger increments, use INCRBY with the desired step.

INCR OFFICIAL_INDEX_VISIT_COUNT

INCRBY OFFICIAL_INDEX_VISIT_COUNT 5

2. Cache Category Tree

Many sites need a category tree. Instead of recursively querying the database on each request, cache the JSON representation in Redis under a key such as MALL_CATEGORY_TREE. The API can then fetch the tree directly from Redis, avoiding multiple database hits. Beware of large keys and consider optimization strategies.

3. Implement Distributed Lock

Redis‑based distributed locks are popular due to their performance. The following Java snippet acquires a lock using SET key value NX PX expireTime. However, this simple approach can release another client’s lock in certain edge cases, so be cautious of pitfalls.

try{  
  String result = jedis.set(lockKey, requestId, "NX", "PX", expireTime);
  if ("OK".equals(result)) {
    return true;
  }
  return false;
} finally {
  unlock(lockKey);
}

4. Build Leaderboard

Leaderboards can be stored with Redis Sorted Set. Use ZADD to add scores and ZRANGE (or ZREVRANGE) to retrieve rankings.

ZADD rank:score 100 "周星驰"
ZADD rank:score 90 "周杰伦"
ZADD rank:score 80 "周润发"
ZRANGE rank:score 0 -1 WITHSCORES

1) "周星驰" 100
2) "周杰伦" 90
3) "周润发" 80

5. Record User Login State

After a successful login, store the user’s session information in Redis with an expiration (e.g., 30 minutes). Subsequent requests check this key to verify the user is logged in.

jedis.set(userId, userInfo, 1800)

6. Rate Limiting

Redis can enforce fine‑grained rate limits, such as allowing an IP address only 10 requests per minute, 50 per ten minutes, or 100 per day. Store a counter per IP (or user ID) and increment it on each request, resetting after a configured TTL.

7. Bitmap Statistics

For tracking daily active users, Redis BITMAP offers an efficient solution. Use SETBIT to mark a user’s login on a specific date and GETBIT to query it. By iterating over dates, you can compute weekly or monthly active user metrics.

SETBIT user:view:2024-01-17 123456 1

GETBIT user:view:2024-01-17 123456

8. Cache Acceleration

Typical cache‑aside pattern: query Redis first; if a hit occurs, return the data. On a miss, fetch from the database, store the result in Redis, then return it. Be aware of cache‑related issues such as penetration, breakdown, and avalanche.

9. Message Queue

Redis supports simple message queues via Pub/Sub. Producers publish to a channel, and consumers subscribed to that channel receive messages. The following Java listener demonstrates handling messages with MessageListener and deduplication using SETNX.

@Slf4j
@Component
public class RedisMessageListenerListener implements MessageListener {
    @Autowired
    private RedisTemplate<String, Object> redisTemplate;

    @Override
    public void onMessage(Message message, byte[] pattern) {
        String channel = new String(pattern);
        RedisSerializer<?> valueSerializer = redisTemplate.getValueSerializer();
        Object deserialize = valueSerializer.deserialize(message.getBody());
        if (deserialize == null) return;
        String md5DigestAsHex = DigestUtils.md5DigestAsHex(deserialize.toString().getBytes(StandardCharsets.UTF_8));
        Boolean result = redisTemplate.opsForValue().setIfAbsent(md5DigestAsHex, "1", 20, TimeUnit.SECONDS);
        if (Boolean.TRUE.equals(result)) {
            log.info("Received result: {}", deserialize.toString());
        } else {
            log.info("Other service processing");
        }
    }
}

10. Generate Global IDs

Redis can generate globally unique IDs using the atomic INCRBY command, which is useful in sharded environments where batches of IDs are needed.

INCRBY userid 10000