How to Track Online Users with Redis Sorted Sets (ZSET) – A Step‑by‑Step Guide

Overview

This guide demonstrates a lightweight real‑time online‑user counter built on Redis sorted sets (ZSET). Four Redis commands are used:

zadd

zrangebyscore

zremrangebyscore

zrem

The sorted set stores each user identifier as a member and the expiration timestamp as the score, allowing constant‑time addition, range queries, and cleanup.

Step 1 – Identify an Online User

For authenticated sites, verify the session token. For public sites, generate a stable identifier (e.g., browser fingerprint) and send it to the backend, typically via a cookie or request header.

import FingerprintJS from '@fingerprintjs/fingerprintjs';

FingerprintJS.load().then(fp => {
  fp.get().then(result => {
    const visitorId = result.visitorId;
    console.log(visitorId);
  });
});

Step 2 – Add or Refresh a User with zadd

Use the current time plus a configurable timeout as the score. This guarantees that a user appears only once in the set because a later zadd overwrites the previous score.

// Java example (Spring RedisTemplate wrapper)
LocalDateTime expireTime = LocalDateTime.now().plusSeconds(expireTimeout);
String expireScore = DateUtil.formatFullTime(expireTime); // e.g. "2024-04-18T12:34:56"
redisService.zadd("user.active", Double.parseDouble(expireScore), userToken);

Step 3 – Query Online Users with zrangebyscore

Retrieve all members whose score is greater than or equal to the current timestamp. The result size equals the number of online users.

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

Step 4 – Periodic Cleanup with zremrangebyscore

Because ZSET entries do not expire automatically, schedule a task that removes all members whose score is less than the current time.

// Java scheduled task
String now = DateUtil.formatFullTime(LocalDateTime.now());
redisService.zremrangeByScore("user.active", "-inf", now);

Step 5 – Remove a User on Logout with zrem

When a user explicitly logs out, delete the corresponding member.

// Java example
redisService.zrem("user.active", userToken);

Key Points

The ZSET key (e.g., user.active) groups all online identifiers.

Member = unique user identifier (token, fingerprint, etc.).

Score = expiration timestamp (epoch or formatted datetime). Using a timestamp ensures that a later login automatically updates the entry.

Adding, querying, and cleaning the set are O(log N) or O(1) operations, making the solution suitable for high‑traffic services.