BiggestBulb,
@BiggestBulb@kbin.social avatar

I think there's quite a bit to say about cost. I'm not sure how Mongo is, but I'm quite familiar with DynamoDB and it gets quite expensive at scale. I'd be worried it'd be unsustainable, especially since postgres is very cheap most of the time (aka, unless you go with RDS).

Also, I'm sure Postgres can be optimized to work well at kbin's scale (since it has worked on sites with millions of users for years). I'm hesitant to say it's an issue with postgres itself, it may be an issue with the queries to the database (maybe the lock?) or even with database auto-scaling (or possible lack thereof).

Also, kbin's general data seems to be highly relational - articles will always have an author, a title, how many votes it has, how many boosts, and all of the comments on it. I think (personally) SQL makes more sense than NoSQL here for that reason, among the others I've listed.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • random
  • wartaberita
  • [email protected]
  • uselessserver093
  • Food
  • aaaaaaacccccccce
  • test
  • CafeMeta
  • testmag
  • MUD
  • RhythmGameZone
  • RSS
  • dabs
  • oklahoma
  • feritale
  • KamenRider
  • Testmaggi
  • KbinCafe
  • Ask_kbincafe
  • TheResearchGuardian
  • Socialism
  • SuperSentai
  • All magazines