We've got some background scripts (notably reapers and format monitors – the ones that are always asking a vendor about specific books) that should be running basically all the time, but actually have significant delays built into them (such as axis_reaper running every four hours and overdrive_format_sweep running once a day).

If those scripts ran to completion all the time, there'd be no issue with the existing timing. There's no reason to complete an Overdrive format sweep more than once a day. The problem is that these scripts usually terminate early, due to a transient API error. Then they don't resume for four hours or a day. This effectively means they run at a fraction of the speed they should be running at.

In the ideal situation, these scripts would run every ten minutes (let's say) and immediately terminate if they'd completed their work recently. That's a little too complicated for cron, so it'd probably be okay for these scripts to run every ten minutes (let's say), period. My main concern with that is small collections would be constantly bugging their vendors about the same books over and over again.