Our FR10 inspection total rose when the monitor reconnected. Same bracket in the fixture, retained completion still high. The proposed patch ignores the first reading after reconnect.
That might hide this duplicate. It also sounds capable of hiding a result completed during the outage. I need a recovery explanation I can hand to operators.
There is a retained job identifier. The local results file also stores it. The monitor's counter ignores that file on startup and just counts observed completion changes.
Reconcile the retained identifier with saved results on reconnect. How does the interface prevent identifier reuse from making a later job look like an old one?
Include both outage cases in your offline tests: a result already saved, and a result produced while the monitor was absent. Ignoring the first reading makes one screen look right by losing evidence from the other.
Identifier reuse is possible after controller maintenance. Controls is revising the interface to include a run boundary. The blanket ignore-first-reading patch is dropped; uncertain matches will stay visibly unresolved during commissioning.