When things calm down, I loaded this link last night off the main 4.20-main report, it claimed test pass rate dropped from 9x.x% down to 0%, then sent me to this link showing no significant regression (green), perma failing in 4.20, 4.19, and 4.18.

test: (note this is not a bug in the test, it is just an example where our test analysis is inconsistent)

[sig-network] network isolation when using a plugin in a mode that isolates namespaces by default should prevent communication between pods in different namespaces on different nodes [Suite:openshift/conformance/parallel]

Main report should match test details for one, but something even fishier seems to be going on with the permafail. Is this test actually passing and we're not seeing the runs? If it is permafail, how did the main report think it was previously 9x.x% passing?