Hmmm...it is possible that you were in the update hole. That is what I'm calling it anyways. The issue stems from the way that we process updates for automatic release. Because of the way that upstream designates their release folders (ie. 7.5.1804, 7.6.1810) the repo location for upstream packages has to change. We manage this in the app-base package. This is why the manual instructions are to run:
What this means is that if you get an app-base, you won't get the next version until the next night. This is why the kernel was semi-pushed when the release happened and had to be withdrawn (some are using it, some aren't). In the future, we will be watching the updates and if app-base is set for update, it will process through a second time.
It is possible that the updates were not processing while it was resolving the two-step delay. I'm speculating here. We couldn't replicate it normally. Let us know if it fails again. We are transitioning the way these updates get distributed so that it follows the yum model more fully.