Every time I post something that gets lots of engagement, I essentially end up carrying out a denial of service attack on myself. What does this say about the design of Mastodon and ActivityPub and the future of the fediverse?
  • @_ed@sopuli.xyz
    8•1 year ago

    Good article, and a new challenge for the network. Once people get to grips with what’s possible Seeing high profile people hosting/posting from their own domains would be great.

    • @Greenpepper@beehaw.orgOP
      6•1 year ago

      Agreed but I also read a message directed to us, the less high profile people that are on large instances. To stop centralization we can decide to move to another, smaller instance. This could be either a Mastodon instance or something like e.g Pleroma or Friendica. This could be either self-hosted or a trust worthy person that owns an existing instance that suits us. Do this in large enough numbers and decentralization will be guaranteed. BTW: here is a link to a list of Mastodon instances and their number of users and statuses: https://instances.social/list

      • @_ed@sopuli.xyz
        3•1 year ago

        Good point - As a hoster of a single-user instance (pleroma) I think more options for managed hosting services could help with this.

  • @nutomicA
    7•1 year ago

    Is this sidekiq software really so inefficient that it can only run 12 parallel jobs on a Mastodon instance that costs 50€ per month? With Lemmy we are running up to 64 such jobs in parallel, and the resource usage is still very low. Part of the reason might be that we are using a Rust library and not a whole separate software project, but still.

    • @Greenpepper@beehaw.orgOP
      6•1 year ago

      I have absolutely no expertise in this area but could it be explained by the fact that Rust is (potentially) much faster than Ruby?

      • @nutomicA
        6•1 year ago

        Yes probably. But what each of these jobs is doing is essentially just creating a cryptographic signature and making an HTTP Post request. So its very fast and jobs can easily be run in parallel. Any modern computer should be able to do this thousands of times per second without a problem.

    • Kromonos
      6•1 year ago

      I already had several issues with sidekiq with GitLab too. It may be a not that bad piece of software, but it looks like, as if it’s overstrained with many requests at once. With GitLab it’s less of a problem, but with software like Mastodon, where there’s a lot going on especially with large instances, it might be worth considering switching to an alternative.