Ok, so a newer fix *has* landed, but a related (possibly key) issue makes it take 6 seconds to count(*) - it will probably be Just Fine on production, but a tweak to the query takes it down to 30ms, so I'm going to go ahead and do that.
Ok, so a newer fix *has* landed, but a related (possibly key) issue makes it take 6 seconds to count(*) - it will probably be Just Fine on production, but a tweak to the query takes it down to 30ms, so I'm going to go ahead and do that.