[Windows] Document workaround to increase ops/sec in single bucket case for append heavy workloads.

Ref: https://docs.google.com/document/d/19l-iDEfM1EetLCoYCjkvNCloVAUfpmIX3PCBOpCx9DU/edit?usp=sharing

It has been observed that there is a drop in ops/sec with append heavy workload on Windows. The workaround is to tune the number of front end memcached threads down to 4 to match the number in 2.5.1. This issue appears to be caused by the scheduling mechanism on Windows. On Linux, with the default number of front end threads the append heavy use case ops/sec has in fact improved.

Linux: http://cbmonitor.sc.couchbase.com/reports/html/?snapshot=ares_251-1083_fec_load&snapshot=ares_302-1518_3af_load
the