There have been several instances where a queued observation gets done twice by RTS2, or does not get received by RTS2. Mike Jones notes a specific instance at: 2022-04-05T07:28:02.402 and 2022-04-05T07:29:59.776.
I am not totally sure if the issue is on the RTS2 or ORP/queue end of things. Creating issue here.