Clustered Redis is coming to Ocim

Discussions - Public
1

Because of the reasons below and more, we’ve decided to start the Clustered Redis epic this sprint. :tada:

  • Open edX no longer supports RabbitMQ, and hasn’t for a long time.
  • Redis 6 supports multi-tenancy, and discoveries have already demonstrated that we should be running a replicated redis service (SE-2487).
  • Stability issues with our RabbitMQ cluster have required lots of SF time during the previous sprint and many prior sprints.
  • Hours were already planned during Sprint 249 to upgrade RabbitMQ and its OS (SE-4540).
  • People in Falcon and Serenity have time available right now.

So @anon46505572 has taken on the epic, and he and @gabor and I are working on tasks this sprint.

If you want to join in, check out our #redis-6 mattermost channel! :grinning:

10 Likes
2

@jill That’s great news! Happy to finally see this project move forward :grinning: :rocket:

Since @anon46505572 is leading it (thanks!) I went ahead and moved the epic from Serenity to Falcon. It is now FAL-2032.

For sustainability (CC @daniel) it would be useful to get a rough idea of the volume of hours that you expect to be logged on SE tickets.

3

Thanks @tikr :) Will discuss this on the epic. TL;DR – it would be great if we could roughly split the hours between Falcon and Serenity. Falcon has more sustainability capacity than Serenity, so it doesn’t have to be even.

1 Like
4

(I meant to post this yesterday but forum was down.)

@jill Sounds good re: continuing the conversation on the epic. I replied there. TL;DR – sustainability aside, Serenity might not have capacity to work on the Clustered Redis epic much in the short term (July).