Improve logging to have indexer logs of around one week

indexer.log is currently covering less than 24 hours as in the log MB-50557. Few top logging statements were removed to avoid more logging. We can further optimise the per vbucket prints by accumulating the state and printing periodically or etc. Prints to StreamBegin, StreamEnd, SystemEvent and etc can use such logging.

Logging this bug to improve logging to have the log data of around one week as requested in MB-50557

Todo:

1. Reduce per vbucket prints in various places
   1. Check if these messages can be converted to stat and dumped periodically
   2. Or If we have some event based logging rather than logging for all vbucket.