Discussed during our office hours SNO team is digging into infrequent failures / pathological events that often correlate with high memory pressure.  Discussion around approaches to see if there were a set of tests that when tested together in parallel caused the condition to occur.

Justin pointed out that the randomization that we currently use inhibits our ability to detect test parings that cause issues and until we have a way to track the seed used (hopefully forthcoming in OTE) we should chose a fixed seed for now and favor the predictability of the test order over the randomization to help identify these issues.

This could help with OCPBUGS-59217 investigation.
Slack Thread