probable regression on 2n-aws mid July 2025

[vrpolakatcisco]

I bet this is an infra issue (AWS changing something on their side) as trending does not show similar decrease on other testbeds. Comparicon tables show all tests are affected, but standard deviation remains high. Not worth bisecting for.