[feature] GoToSocial multinode / cluster support

[NyaaaWhatsUpDoc]

Multinode / clustered scaling support for GoToSocial has been requested several times during the current (alpha) stage of development. While we understand that you want to be able to run it in an easily scaled / redundant way, and appreciate your enthusiasm (really, it keeps us going!), we will not be implementing (or accepting pull requests with) multinode / cluster support for the foreseeable future.

There are several reasons for this:

• some basic features are still missing (e.g. lists, tags)
• GoToSocial will always prioritize scaling down (think, your toaster!)

Finally, and most importantly, there are pressing safety / anti-harassment issues that need to be addressed, including the following:

• restricted (approval-based) registration
• subscribable block lists https://github.com/superseriousbusiness/gotosocial/issues/3
• anti-spam tooling
• improved moderation tooling
  • e.g. lock-down accounts of users during mental-health crisis (timelines invisible but DMs still work)
  • e.g. improved reporting system with (GDPR compliant) event archival tools (must support resolutions)
  • e.g. automatic disabling of registration until certain active-users to moderators ratio is hit
• further discussions regarding combating fraud / abuse / spam at scale on federated networks
• further discussions regarding combating / disrupting right-wing radicalization pipelines

The above is not intended to be an exhaustive list, nor a task-list of "blockers" standing in the way of multinode support. Instead this provides context as to why we're not (yet) interested in scaling GoToSocial upwards, and a good indication of what protections should at least be implemented before we will even consider it. Ideally, the more protections, the better!

The harm that social media has caused in the past decade(s) has been devastating, due among other things to greed, carelessness, and a lack of interest in addressing basic safety and moderation concerns like the above. We (the GoToSocial team) have no intention of releasing yet another attack vector into the world (especially one that is capable of large-scale operation) without thinking carefully about harm mitigation.

Thank you for reading!

---

Also general note from kim -- I've had the above moderation tooling improvement suggestions rattling around in my brain for months now so it feels good to get them down on paper (silicon).

[NyaaaWhatsUpDoc]

it's been a while since I wrote this, I'll write an update for this over the next few days. (our opinion on this still has not changed)