kyverno: local: scale down deployment and stop requesting CPU/MEM
Don't request any particular CPU/MEM, which should help the lima-kilo k8s scheduler don't pre-allocate CPU/MEM for kyverno, which is mostly idle anyway.
Also, scale down the deployments wrt. replicas.
Signed-off-by: Arturo Borrero Gonzalez aborrero@wikimedia.org