HenryAllen0264 · 2 May 2025, 01:44 UTC
Can I ask whether the receiver retains an attempt identity or result you can query? We had the same uncertainty around sending in our own bookkeeping review, and another local status alone could not tell us what reached the other side.
20 pointsNo query available. It returns a result while connected, then our app saves it.
10 pointsThen keep those interrupted attempts held for a defined review while you agree how to reconcile them. The receiver may have completed the inspection, and a second send could repeat work rather than recover the missing result.
23 pointsWhat does the operator see after restart? Pending sounds like ordinary queued work to me. I would not know it meant somebody has to investigate.
-8 pointsCurrently pending. Proposed label is outcome unknown, with automatic replay disabled for those entries.
10 pointsHow will the application distinguish never submitted from a crash during dispatch? That's the boundary I'd test before relying on the new label.
17 pointsDeveloper proposes a durable dispatch-intent entry before sending. Incomplete attempts after that point stay unknown.
6 pointsWho can resolve an unknown attempt, and what evidence must they check before choosing any recovery action?
15 pointsTest a crash after intent is saved but before the actual send too. That record will be uncertain locally even if nothing reached the receiver; the recovery procedure must allow for both possibilities.
18 pointsWe had an operator interpret unknown as failed and repeat a check. Put the reason for the hold beside the label so that distinction does not depend on remembering a training talk
13 pointsShift lead will own review with the integrator's recovery procedure. The procedure is not finished yet.
19 pointsWill the screen show which bracket and attempt need review? A general warning would leave me hunting through the queue.
17 pointsYes. Bracket identity, attempt, last saved stage and review owner in the proposed view.
17 pointsDid the developer run the crash cases offline? Include a received result lost before local saving, since that is another route to a completed physical inspection with an uncertain local record.
15 pointsDoes the restart also retain decisions already saved by the reviewer?
12 pointsThat is worth checking as a separate saved update, especially if the review screen and inspection-event handler write the same record.
8 pointsOur team found a short walkthrough with an operator helpful here; they asked what happens when the named reviewer is absent, which the developer had not considered
16 pointsOffline crash tests hold uncertain attempts and retain saved review decisions. No production change installed yet.
8 pointsBackup review owner added. Receiver history and the recovery procedure still need agreement before deployment.
10 pointsPlease bring back the finished screen wording. I would like to see whether someone on shift can understand the hold without reading the software notes first.
9 pointsThis discussion is closed to new replies after six months without activity. Last activity: .