I'm seeing this on multiple fresh 1.5.0 installs https://github.com/harvester/harvester/issues/8004

Agree. But also I’m finding that on the “master” node, of a 2 node bare-metal cluster, which is running no VMs - most of the CPU is reserved, leaving virtually nothing for VMs.

Reserved 5.42 of 7.02 C */ 77%*