We noticed in a HELP ticket that secondary update performance is slower in 5.0 compared to primary updates for a workload with single writer targeting the same document using batched retryable write workload.

Extra notes:
1) Customer was running 5.0.3. We improved the update performance in 5.0.22 (SERVER-60176). I still need to confirm if there are any other opportunities that can improve update performance on secondary for the workload the customer was running?
2) In terms of performance benchmarks so far, I observed that version 4.4 performs better than 5.0.26 and Master for batched retryable update targeting a single document. Master's performance varies depending on the document size.