In some customer cases using dbCheck, a common recommendation is to throttle dbCheck by setting batchWriteConcern: n = # of nodes in the set. This is to pause the next dbCheck batch until the previous batch has been replicated and applied on all nodes. However, now that we have separated lastWritten and lastApplied in v8.0+, we should clarify to what extent batchWriteConcern will be successful in throttling dbCheck.

Acceptance criteria:

• Doc explaining the current implementation (code dive)
• Playbook update for this technique