Spurious kubediff due to CPU resources

[jml]

We are currently seeing this kubediff on prod:

## scope/pipe (Deployment)

.spec.template.spec.containers: 'element [0]' missing
## monitoring/alertmanager (Deployment)

.spec.template.spec.containers: 'element [0]' missing

AFAICT, this is because for scope/pipe, the YAML file says:

        resources:
          requests:
            cpu: 0.5

But the k8s object says:

        resources:
          requests:
            cpu: 500m

Despite being freshly applied.

Since this difference isn't meaningful, it shouldn't be reported.