Elegant Ways to Implement Online User Counting with Redis ZSET

Introduction

Online user counting is a common requirement; many implementations exist. This article shows a practical approach using Redis sorted sets (ZSET).

Implementation Steps

1. Determining Whether a User Is Online

If the site requires login, the validity of the user's token indicates online status. For public sites, a rule such as browser fingerprint can be used. Browser fingerprint may include User-Agent, HTTP headers, screen resolution, timezone, language, plugins, etc. Libraries like FingerprintJS or ClientJS simplify generating a unique identifier.

import FingerprintJS from '@fingerprintjs/fingerprintjs';
// Initialize
FingerprintJS.load().then(fp => {
  // Get visitor ID
  fp.get().then(result => {
    const visitorId = result.visitorId;
    console.log(visitorId);
  });
});

The identifier is stored in a cookie or header and sent to the backend, where it serves as the member of the ZSET.

2. Adding an Online User with ZADD

The ZADD command takes three arguments: key, score, member. Example:

ZADD myzset 1 "one"

In the implementation, the score is the expiration timestamp (current time plus a timeout). Adding a token:

// expireTime = now + expireTimeout seconds
String expireTimeStr = DateUtil.formatFullTime(expireTime);
redisService.zadd("user.active", Double.parseDouble(expireTimeStr), userToken);

Because the score is unique per login, repeated logins replace the previous entry, ensuring a single latest record per user.

3. Querying Online Users with ZRANGEBYSCORE

ZRANGEBYSCORE retrieves members whose scores fall within a range. Example:

ZRANGEBYSCORE myzset 1 3

To get all currently online users, query from the current timestamp to +inf:

String now = DateUtil.formatFullTime(LocalDateTime.now());
Set<String> userOnlineStringSet = redisService.zrangeByScore("user.active", now, "+inf");
int onlineCount = userOnlineStringSet.size();

4. Periodic Cleanup with ZREMRANGEBYSCORE

ZREMRANGEBYSCORE removes members whose scores lie in a given range. Example:

ZREMRANGEBYSCORE myzset 1 3

A scheduled task deletes entries whose expiration time is earlier than now:

String now = DateUtil.formatFullTime(LocalDateTime.now());
// Remove entries with score between -inf and now
redisService.zremrangeByScore("user.active", "-inf", now);

5. Removing a User on Logout with ZREM

ZREM deletes specific members from the ZSET. Example:

ZREM myzset "xxx"

When a user logs out, the backend calls:

redisService.zrem("user.active", "xxx");

Conclusion

The core idea is to maintain a ZSET where the key represents the online‑user set, each member is a unique user identifier, and the score is the expiration timestamp. Adding, querying, and cleaning the set provides a simple yet effective online‑user counter.