Use NamespacedName Type for Internal APIs

[danehans]

Internal APIs, e.g. EndpointSliceReconciler, use separate ns and name strings to represent the ns/name of CRs. A few benefits of using NamespacedName are:

• Encapsulates the namespace and name into a single struct, improving readability by clearly indicating that these two fields together uniquely identify a Kubernetes resource.
• Provides stronger type safety compared to using two separate strings.
• Provides built-in methods like String() that automatically format the resource identifier in the conventional namespace/name format, reducing boilerplate code and potential formatting errors.
• Functions that operate on Kubernetes resources often require both namespace and name. Using a single types.NamespacedName reduces the number of function parameters, simplifying function signatures.

[ahg-g]

and avoids switching the two parameters, which will likely be placed after each other in func parameters

[danehans]

The inferncepool and inferencemodel reconcilers are using separate string fields instead of NamespacedName even with #165.

/reopen

[k8s-ci-robot]

@danehans: Reopened this issue.

In response to this:

The inferncepool and inferencemodel reconcilers are using separate string fields instead of NamespacedName even with #165.

/reopen

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[MadhavJivrajani]

Apologies @danehans, I've created #172 to fix this