suddenly got much worse in the past week. [sig-api-machinery] Kubernetes APIs remain available past 7 days 7-14 days all 74.42% 86.78% aws 81.82% 100.00% azure 57.14% 60.98% gcp 81.82% 100.00%
It appears that test-grid data is incorrect. This incorrectness is reflected in sippy, which uses the test-grid summary of job runs and test runs to determine pass fail rates. See https://testgrid.k8s.io/redhat-openshift-ocp-release-4.6-informing#release-openshift-origin-installer-e2e-gcp-upgrade-4.5-stable-to-4.6-ci and the runs on 10/10 that show several failed tests. Those tests are actually flakes, not failures.
I was also going to point this out, thanks for the comment on the data being incorrect David.