/// Check whether to save state to db every n sends if there's no failures (during failures state is saved after every attempt)
/// This determines the batch size for loop_batch. After a batch ends and SAVE_STATE_EVERY_TIME has passed, the federation_queue_state is updated in the DB.
staticCHECK_SAVE_STATE_EVERY_IT: i64=100;
/// Save state to db after this time has passed since the last state (so if the server crashes or is SIGKILLed, less than X seconds of activities are resent)
/// interval with which new additions to community_followers are queried.
///
/// The first time some user on an instance follows a specific remote community (or, more precisely: the first time a (followed_community_id, follower_inbox_url) tuple appears),
/// this delay limits the maximum time until the follow actually results in activities from that community id being sent to that inbox url.
/// This delay currently needs to not be too small because the DB load is currently fairly high because of the current structure of storing inboxes for every person, not having a separate list of shared_inboxes, and the architecture of having every instance queue be fully separate.
/// (see https://github.com/LemmyNet/lemmy/issues/3958)