WT-11941 is (likely) going to add some assertions around the time recorded for reconciliation, building a page image, and history store reconciliation. However, asserting rec_hs_wrapup and rec_img_build takes less than a day (similar to the rec statistic) fails intermittently.

The with rec_hs_wrapup is that the hs_wrapup_start field in the reconcile_timeline can be zero (with a finish time of non-zero) by the time we get to wt_reconcile. The code looks obvious, so it's possible this is another re-entrancy problem like WT-11941.

I haven't looked into the rec_img_build assertion failure, but I imagine the problem is similar.