This seems to be hitting gcp the hardest and blocking payloads on failed installs where operators are not ready (yet are by the time we go collect artifacts), upgrades taking more than allotted time in the tests, and slow OSUpdateStaged events.

Tests affected that we know of:

install should succeed: cluster creation

install should succeed: overall

[sig-cluster-lifecycle] cluster upgrade should complete in a reasonable time

[bz-Machine Config Operator] Nodes should reach OSUpdateStaged in a timely fashion

However component readiness has degraded badly since and it's possible this is affecting many more tests than we can immediately correlate.

Tentative explanation is a change to quay image caching by testplatform, moving to R2, which is being adjusted now. See this thread for details