Our FR10 fixture-check monitor counts a completion again on reconnect although the indication was already high. How should I match it to its original attempt without clearing it or resending work?
Check the retained identity and result beside the processing history that survives restart. The first high state after reconnect is an observation of retained data, not proof that a new job just completed.
The retained record has a unique attempt identity and coupon link. Our saved detail row already contains that result once, but the displayed total increments when the monitor sees high after connecting.
Does the total mean completed attempts or accepted coupons? The same stored result can support a technical completion count without establishing the coupon's current quality disposition.
Anna, this screen explicitly reports completed attempts; quality dispositions are elsewhere. Zara, the replay rebuilds the expected total from distinct durable attempt rows. We have stopped incrementing a separate number just because the retained bit is high.
Include a completion first discovered after disconnection too. Reconnect must not add a duplicate, but it still needs to record genuinely new identified results that arrived while the monitor was absent.
Both pass in the replay: repeated retained result stays one row, and a newly discovered identified completion adds one row and one completed attempt. Reopening rebuilds the same total. The old standalone counter is no longer read.
Have those tests used the actual retained-record format from your documented interface? A convenient fake identity would test the counter while leaving the real input path unexamined.
Yes, saved records from the installed interface, then the maintainer repeated the reconnect comparison through the normal monitor. Counts match the durable rows after reopening. Thanks Anna; the screen label remains completed attempts, not accepted coupons.
Keep that distinction and the newly discovered-result test in the handover. The fix now accounts for both replayed history and work first observed after reconnect, without changing the job's physical state.