From debugging the WT-8092, I have found that the cursor_next_skip_lt_100 and cursor_next_skip_ge_100 combines the increments of both history store, WT data files. When performing a search_near() for a particular key without any visibility in a populated btree, the __wt_txn_read function checks the if a particular entry is valid to be returned. The visibility of the entry determines if an entry is valid. The function will first check inside the update list, ondisk, and finally the history store for any visible entries. Now when we are looking inside the history store another search near is started to now track if there exists an entry in the history store. Now the cursor_next_skip_lt_100 and cursor_next_skip_ge_100 gets incremented because we have done a search near inside the HS. This means to gather valuable information around number of entries traversed through a single search function can't be determined whether if it was from the history store or the data file.

The idea behind this ticket is to potentially add another statistic so that we can increment statistics seperately from the HS. This benefits WT to better diagnose if we are tracking next calls from a data file or HS file.