Handling Dynamic Group Membership with Kafka and Node.js - Stack Overflow

admin管理员组

文章数量:1279089

I'm developing a chat application using Kafka (with KafkaJS) and WebSockets in Node.js. Users can dynamically join and leave chat groups, sometimes multiple times within short periods (online and offline). I need to ensure each user receives only those messages sent to the group during the time they were actually members. Specifically, I'm trying to handle these edge cases:

User is added to a group while offline and should only receive messages sent after the addition.

User is removed from a group while offline and should only receive messages up until removal.

User is added to a group while online and should immediately start receiving new messages.

User is removed from a group while online and should instantly stop receiving new messages.

User repeatedly joins and leaves a group (offline), and upon reconnection, they should receive messages only from the periods they were members.

Currently, I maintain membership events (added/removed) with timestamps in Redis, and each Kafka message has a timestamp. My Kafka consumer subscribes to relevant topics, but I'm unsure of the best way to filter and deliver messages accurately based on these dynamic membership windows.

How can I efficiently filter Kafka messages to ensure users receive only the messages sent during their actual membership periods, especially given rapid or frequent changes in group membership?

本文标签： Handling Dynamic Group Membership with Kafka and NodejsStack Overflow