All RHEL systems installed using Anaconda get a resume=xxx parameter.
It appears that if the device specified is not found, the boot hangs forever waiting for the device to show up, e.g.:

Job xxx.device/start running (1min 42s / no limit)

IMHO there should be a timeout associated to this (e.g. the standard device timeout).

Typical use case is there is resume=/dev/md/swap (Swap on Software Raid) configured and the device is not found, because the initramfs being used for booting is a generic initramfs and not host-specific one.
In such case, mdadm (or probably more udev rules) will generate /dev/md/<hostname>:swap device node, instead of /dev/md/swap.

Another use case is having resume=/dev/md/swap and upgrading the system using leapp, which generated a host-only initramfs with mdadm configuration explicitly excluded from the initramfs.