aws_eks: containerPort and apiVersion should not be in snake_case (i.e. container_port and api_version)

### link to reference doc page

https://docs.aws.amazon.com/cdk/api/v2/python/aws_cdk.aws_eks/README.html#quick-start

### Describe your issue?

When going through the aws_eks quickstart as-written in Python, the deploy fails with errors like this:

```
4:41:19 PM | CREATE_FAILED | Custom::AWSCDK-EKS-KubernetesResource | helloeksmanifestmypodC48C8B9D
Received response status [FAILED] from custom resource. Message returned: Error: b'error: error validating "/tmp/manifest.yaml": error validating data: [ValidationError(Pod.spec.containers[0].ports[0]): unknown field "container_port" in io.k8s.api.core.v1.ContainerPort,
ValidationError(Pod.spec.containers[0].ports[0]): missing required field "containerPort" in io.k8s.api.core.v1.ContainerPort]; if you choose to ignore these errors, turn validation off with --validate=false\n'
```

(The specifics vary slightly for the apiVersion and containerPort cases. That's from the containerPort issue.)

I'm guessing that these are converted to snake_case somewhere in the documentation build process (e.g. whatever converts addManifest to add_manifest), but these particular keys should be the same as in the TS version.

Start with the linked AWS CDK EKS Python quick-start page and compare its containerPort and apiVersion keys with the TypeScript example. Inspect the documentation source or generation path that produces the Python example. Done means the as-written Python quick-start deploys successfully without converting these Kubernetes keys to snake_case.

Written by the indexing model from the issue text.