SQL Server 2017

By using db<>fiddle, you agree to license everything you submit by Creative Commons CC0.
select v.*
into works
from (values (1, convert(date, '2019/04/20'), convert(time, '08:30'), convert(time, '14:20')),
             (2, '2019/04/21', '09:05', '15:15'),
             (3, '2019/04/22', '08:37', '14:22'),
             (4, '2019/04/23', '08:05', '15:00')
     ) v(d, Date, WorkStartTime, WorkEndTime)
4 rows affected

select v.*
into leaves
from (values (1, convert(date, '2019/04/20'), convert(time, '08:00'), convert(time, '9:00')),
             (2, '2019/04/21', '08:30', '16:00'),
             (3, '2019/04/22', '07:00', '08:45'),
             (4, '2019/04/22', '14:20', '17:00')
     ) v(d, Date, LeaveStartTime, LeaveEndTime)
     
4 rows affected

select l.date,
       sum(datediff(minute, l.LeaveStartTime,
                    (case when l.LeaveStartTime < w.workStartTime then w.workStartTime else l.LeaveStartTime end)
                   ) +
           datediff(minute,
                    (case when l.LeaveEndTime > w.workEndTime then w.workEndTime else l.LeaveEndTime end),
                    l.LeaveEndTime
                   )
          ) as overlaps
from (select l.d, l.date,
             (case when LeaveStartTime < '08:00' then '08:00' else LeaveStartTime end) as LeaveStartTime,
             (case when LeaveEndTime > '16:00' then '16:00' else LeaveEndTime end) as LeaveEndTime
      from leaves l
      where LeaveEndTime > '08:00' and LeaveStartTime < '16:00'
     ) l join
     works w
     on w.date = l.date and
        (l.LeaveStartTime < w.WorkStartTime or
         l.leaveEndTime > w.workEndTime
        )
group by l.date


date
overlaps


20/04/2019 00:00:00
30

21/04/2019 00:00:00
80

22/04/2019 00:00:00
135
date	overlaps
20/04/2019 00:00:00	30
21/04/2019 00:00:00	80
22/04/2019 00:00:00	135