You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Inspired by what fediverse.network does to detect spam instances:
group instances by "topmost registrable domain", e.g. foo.example.com and bar.example.com get grouped by "example.com";
every day, check only one instance of the group. Or, in other words, the check interval for an instance would be number_of_instances_in_a_group*day.
As a result, a spammer that created a million instances won't affect a spider's load at all, as only one of the million instances will be checked every day. It still affects the databse size though, and adds overhead in detecting groups.
The text was updated successfully, but these errors were encountered:
Inspired by what fediverse.network does to detect spam instances:
As a result, a spammer that created a million instances won't affect a spider's load at all, as only one of the million instances will be checked every day. It still affects the databse size though, and adds overhead in detecting groups.
The text was updated successfully, but these errors were encountered: