Excessive flushes may kill performance on complicated pipelines

Observed in CBSE-14959, Analytics can get overwhelmed with frequent flushes combined with complicated pipelines. In this case, there were 59 secondary indexes & three datasets all defined on a single collection. The excess flushes were caused by:

• flushes forced due to incorrectly perceived uuid updates
• regression introduced with fix for MB-57315, resulting in perpetual flushes on an overloaded system (one likely overloaded due to flushes)