I don’t know how Lemmy’s codebase works but I would imagine it’d be something like the server detects it’s under heavy load by its CPU usage, sends a request out to a given number of its friends, and they respond back whether they have spare capacity, and if one says it’s got capacity it redirects users to their domain instead (or maybe asks them if they want to first and then redirects them)
Could even just have a pop up saying “This instance is overloaded, please consider using one of the following:” and then give a list of randomly selected instances with free capacity with links to continue with wherever you were doing on the other instance
Problem with this is that every time this happens one needs to research those instances because probably most of the people wouldn’t like to end up in beehaw situation. But I guess it is only matters if this overloads would happen frequently. Once every few months I wouldn’t mind changing instance if it improves overall performance, and makes site more usable for everyone.
Beehaw defederated from loads of instances right? Kinda saw that coming seeing how much they’d setup on their own instance and branded everything
This load balancer could prioritise instances that have similar white/black lists to the one you’re current on, and if it suggested switching instead of doing it transparently people would be choosing which one of those they’d rather use
I think it’d also help people better understand how the system works and maybe permanently switch to another less popular instance
interesting concept, i wonder how that would be implemented
I don’t know how Lemmy’s codebase works but I would imagine it’d be something like the server detects it’s under heavy load by its CPU usage, sends a request out to a given number of its friends, and they respond back whether they have spare capacity, and if one says it’s got capacity it redirects users to their domain instead (or maybe asks them if they want to first and then redirects them)
Could even just have a pop up saying “This instance is overloaded, please consider using one of the following:” and then give a list of randomly selected instances with free capacity with links to continue with wherever you were doing on the other instance
Problem with this is that every time this happens one needs to research those instances because probably most of the people wouldn’t like to end up in beehaw situation. But I guess it is only matters if this overloads would happen frequently. Once every few months I wouldn’t mind changing instance if it improves overall performance, and makes site more usable for everyone.
Beehaw defederated from loads of instances right? Kinda saw that coming seeing how much they’d setup on their own instance and branded everything
This load balancer could prioritise instances that have similar white/black lists to the one you’re current on, and if it suggested switching instead of doing it transparently people would be choosing which one of those they’d rather use
I think it’d also help people better understand how the system works and maybe permanently switch to another less popular instance