Cannot match this UR5e acknowledgement to either plate-check attempt

Yes, startup line is present. Two runs, same local job label in each. I've marked the boundary.

Keep each run's event order intact. Don't let sorting timestamps erase that boundary.

Original capture requested. Current export has the job label and acknowledgement state, no request counter.

Also establish what the acknowledgement means in this interface, whether it confirms receipt or a completed inspection; the name alone leaves room for a second wrong assumption.

Luis, operator remembers the plate still in the fixture but cannot say whether a check happened. Anika, original files are preserved. Jamie, acknowledgement means request accepted, with inspection result reported separately.

Then keep that result question separate. Accepted request isn't an accepted plate.

I have removed the report's link from acknowledgement to passed inspection. No result is matched to either attempt yet.

Did controls find an additional sequence field in the original capture, Ella? That could still help with attempt ownership even though it won't supply the missing inspection result.

There is a counter, but capture begins after the restart and nobody has its earlier value. Same job label still does not identify which request produced the retained acknowledgement.

Does the counter persist across that kind of restart? I would want controls to confirm that before interpreting the later value, even if the earlier one is unavailable.

And don't turn the operator's remembered plate position into a completion check; ours stayed in the fixture for both failed and successful attempts.

Understood, Anika. That observation stays as location only. Controls is confirming counter behaviour; I haven't used it to draw a new arrow.

Who owns the unresolved plate disposition while you investigate?

Quality owns it. They have left this plate unresolved, separate from work on future logging. Controls confirms the counter survives the application restart, but the missing earlier value still prevents the comparison we need.

For future attempts, agree an identity that both sides retain and include it in the saved result; preserve startup boundaries too, rather than relying on matching wall clocks later.

Those changes are being specified. Old acknowledgement remains unassigned. Thanks Daniel for starting with the restart boundary; it stopped the first report from looking much more certain than our evidence.