Looking at the performance dashboard (links and screenshots attached/commented) there's a performance regression that hovers around 25% for the following tests:

• medium-btree - Load Time
• evict-btree - Load Time
• overflow-130k Btree - Load time

Additionally, there's a ~40% regression in "large record btree throughput - update, modify and modify force update".

All of the above charts implicate commit #22,694, which is a1f21fe61929c12b360d8b50f03bc1584d7b10fd (WT-9310). The last known-good is commit #22,692, which is 1696bc595ca30ebeb0c2ecd5170c60b2367c83d2 (WT-9479).

Scope for this ticket is to investigate further and accurately identify the offending commit. #22,694 would be the first thing to check, but if that change is (e.g.) documentation or comments, then we need to dig a little further.