openshift-sdn was using JSON-only k8s clients because earlier releases did not support using protobuf clients to work with CRD types. This is apparently fixed in k8s 1.16. I don't know if it's worth bothering testing the fix beyond "things still work" but if so, Clayton can fill in the expected difference in metrics output.
We were running during an e2e test about ~1.4-1.8 cores for all SDN containers across all 6 nodes. We are now running about 1.1 to 1.4 cores. In heavily used clusters this will be more impactful.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:0062