Our FR5 dashboard has more passes than accepted housings

Physical housing ID is present on each row, with an attempt number; the dashboard simply counts rows marked pass.

Then build the accepted-item total from the agreed item disposition, while retaining the linked attempts. Don't replace the row count with any-pass-ever unless quality says that is actually the acceptance rule.

What happens when a housing passes, then a later reviewed check fails? That will show whether the proposed total follows the decision you really use.

Try that case in a rebuilt report first. Also reopen it after restart; totals should come from the stored decisions, not another round of increments.

Quality's current disposition controls release, Grace, not any historical pass. We found one housing whose later failure was hidden by the dashboard's cheerful count.

Who has that housing now while the count is corrected?

Quality has it on hold; it was not dispatched, and the trial report now shows each housing once with attempts underneath and its current disposition beside it.

Good that it stayed put. Does the report distinguish held from rejected? Both may be excluded from accepted quantity, but they are different things for the person arranging the next work.

And don't let held items fall into an automatic retry list. We just found that mistake behind a perfectly clear hold tag elsewhere.

Held and rejected are separate now; reopening the offline report keeps the same total, but we still need to test later disposition changes and restart against the installed system.