Make the explanation of the Medusa event bus, Redis and BullMQ more explicit

[mateomoreno]

What Medusa version and documentation are you using?

v2

Preliminary Checks

• This issue is not a duplicate. Before opening a new issue, please search existing issues: https://github.com/medusajs/medusa/issues

Issue Summary

Our Redis instance spiked using 3.82GB of memory with over 1.3M keys. Investigation showed that 1.29M keys (99%) are RedisEventBusService entries that have accumulated over the past 6 weeks with no cleanup. We thought this is because we had a scheduled job that ran daily at 3 AM, emitting vendor.pricing.update events for batch processing (~30,000 events/day). Without a retention policy, these events were not getting purged from Redis. We were under the impression all redis events had a ttl by default. However, upon understanding BullMQ and its relationship with Redis, we realized that BullMQ's default is removeOnComplete = true. But that doesn't actually mean events are deleted it just marks them as eligible to be deleted. Without {age, count} object, lazy cleanup never triggers.

How can this issue be resolved?

1. We added this to our medusa-config.ts

{ resolve: '@medusajs/medusa/event-bus-redis', options: { redisUrl: process.env.REDIS_URL, jobOptions: { removeOnComplete: { age: 3600, // Remove completed jobs after 1 hour count: 1000, // Keep max 1000 completed jobs }, removeOnFail: { age: 86400, // Remove failed jobs after 24 hours count: 5000, // Keep max 5000 failed jobs }, }, }, },

Are you interested in working on this issue?

• I would like to fix this issue

[shahednasser]

Thanks for opening this issue! Your suggestion is valid, I'll look into documenting it