@evan @debirdify The system that does the backups is distinct from the servers handling normal user traffic. If you want your archive get it while you can.
@evan The queue item is the request, not a batch of tweets (subset) from the request. A worker will complete a user's archive before moving on to the next.
@richunger I don't think you understand my point.
Clogging the queue when you already have a mostly-complete backup is unfair to the people who have no backup.
They may not get any back up at all before the wheels come off the cart.
@evan oh I see what you mean
@richunger usually this kind of thing runs with a queue that's handled by a limited number of job-handling servers. The more people in the queue, the longer it takes to get your backup.
Maybe that's not how it works at Twitter.
But if it is, people who already have 99% of their stuff backed up might want to wait to let people with 0% of their stuff backed up get a backup before things go 100% sideways.
If you're aware of another structure that makes this unnecessary, lmk.