I'm reconstructing a fixture-inspection timeout, but the Python, PLC and UR5e clocks disagree and the app reused the housing job label after restart; what can I establish without assigning the acknowledgement to whichever timestamp looks closest?
Preserve the original records and separate the application runs at restart. Look for request or attempt identifiers, recorded sequence and message content before comparing clock times. A common label may describe the housing job without uniquely identifying the attempt that received an acknowledgement.
I can separate the application runs, but neither acknowledgement nor request includes a persistent attempt identifier and I have no measured clock offsets for that period.
Then preserve the ambiguity. You may be able to bound possible associations from ordering within each source, but cross-system proximity is not proof. Also distinguish receipt acknowledgement from inspection completion; even a correctly assigned receipt would not establish the housing's result.
I've recorded both possible associations and left the acknowledgement unassigned; quality's housing disposition is a separate enquiry, not a result inferred from this timeout.
For future attempts, ask the application and interface owners for identity that survives restart and is carried through the relevant records. Record clock relationships too, but don't expect that improvement to retrospectively decide this old event.