This ticket is similar in idea to SERVER-89948 and SERVER-78834. The current oplog application code can be difficult to parse through at times. Some specific points of improvement to consider:

1. Functions like this one have complicated `if` logic that can be difficult to trace. There are a lot of conditions to consider when reading through the code
2. Similar to above, these functions are also long, with a lot of tabbing
3. Errors that are intended to be thrown are caught and returned instead, but then later thrown higher in the stack. LOGV2_FATAL_CONTINUE seems like a scary paradigm to use if we do not immediately throw afterwards