The or_recordid_deduplication workload currently does not cause any unavailability as measured by the Atlas read/write canary.

This ticket is to determine if:

• The workload is working as intended on the Atlas variants
• The workload needs any additional tuning to reproduce the availability issue on Atlas

If it is the intended behavior for the workload not to cause unavailability as measured by the canary, either due to the type of issue it is reproducing or because the issue is already fixed on master, feel free to close this ticket with a corresponding explanation.