Apparently witches.town going offline can potentially affect instance performance.
Not sure if it will affect social.coop, but this toot suggests we might need to stop Sidekiq contacting witches.town.
When witches.town goes down in the next few days, you may notice Mastodon running more slowly, as Sidekiq workers try to contact it and have to wait for the request to time out before they can move on to the next job.
To handle this, AFTER w.t goes down, log into your server on the back end, run redis-cli, and enter:
sadd unavailable_inboxes https://witches.town/inbox
This tells Mastodon not to bother trying to talk to witches.town.