[RFE] Count only days of service for tallying broken hosts mid-month before inoperable

[sadsfae]

This is related to https://github.com/redhat-performance/quads/issues/390 so would likely come after that is resolved, which would not count broken systems as being idle/unused in our monthly projected percent utilization reports. For now (covered in #390) we should simply list broken systems as utilized.

This RFE covers considerations where systems have been in active assignment service during a given month but is subsequently marked broken mid-month. In these frequent scenarios we lose out on this level of accuracy in reporting.

Suggestion

• Count the days of the month up until the system is marked broken.
• Utilize strike-through or some clear visual inicator on the days following when the system is rendered inoperable and "broken"
• Still count the remaining days a system is marked broken towards utilization as per #390

Implementation

• This would likely require recording timestamps of when a system is marked broken and marked repaired to be wholly accurate.
• Having broken/repaired timestamps may feed into more detailed reporting later e.g. how long was a set of systems actually broken? etc.