Queue normally shows an average wait time over the last hour. The main problem is that an average over the last hour, for a lab that takes 2-4 hours, is almost always off from what students truly need to wait. Instead, a better predictor for wait time is the time the oldest pending request has been waiting for.
This MR adds this current wait time. It is defined as the wait time of the oldest pending request. If no such request exists, it uses the wait time that was experienced for the last handled request.
The calculation of the current wait time also works with timeslotted labs. Here the start of the timeslot is used rather than the creation time of the request.
Of course, current wait time is also not perfect. Because it is not an average, the value is expected to fluctuate a lot more. Prioritization of requests also messes with the current wait time. Therefore, both values are displayed.
Taico Aerts <t.v.aerts@tudelft.nl>
