I matched the housing-stop video and export by attempt before sending this UR5e ticket. Support now says the fixture confirmation may change between the exported rows. Our file is a sampled summary, not the underlying event record. Which version should the recorder owner provide so we can examine that moment?
Ask for the underlying state changes over the same identified attempt, with the recorder's timing and coverage stated. The sampled file can be useful for trends without showing every short event. Have the owner confirm what was actually retained before promising support a finer trace
The raw event record was retained. Recorder owner found the export had reduced it to regular snapshots, leaving a brief confirmation change out. They've supplied the original events for this attempt with the sampling note attached to the old file.
Check that the event names use the installed fixture mapping too. You fixed the interval and coverage; no need to let an old alias turn the newly visible event into a different physical input
Mapping checked against the installed fixture and job. The additional event is its seated confirmation falling before the application wait. Support has that identified sequence rather than the old summary claiming it stayed true throughout.
Their engineer identified the branch waiting for a renewed fixture exchange confirmation after the loss. Maintenance is checking why the seated state dropped. The recorder owner also tested the revised fault export with a retained short event; it now preserves that event and its attempt identity.
That gives maintenance a specific physical question. Keep the export fix separate from the fixture fault in the ticket, so nobody closes the handling stop merely because the next file is more informative