[EKS] [Logging]: Add ability to configure control plane log level (TRACE/DEBUG/INFO/WARN/ERROR)

## Feature Request

### Summary

Amazon EKS does not currently expose a way to configure the log level for control plane components. Users have no control over log verbosity, which has both cost and operational implications.

### Problem Statement

EKS control plane logs are streamed to CloudWatch at a fixed verbosity level. Without the ability to tune log levels, users face two pain points:

1. **Cost**: In stable production environments, verbose logs (DEBUG/TRACE) generate unnecessary log volume, driving up CloudWatch ingestion and storage costs. Most of the time, `WARN` or `ERROR` level is sufficient.
2. **Troubleshooting**: When an issue does arise, there should be a way to temporarily raise the log level to `DEBUG` or `TRACE` to get the detail needed for diagnosis — without recreating the cluster or relying on AWS Support.

### Use Cases

- **Normal operations**: Run at `WARN`/`ERROR` to minimize CloudWatch log volume and reduce cost.
- **Active troubleshooting**: Temporarily elevate to `DEBUG` or `TRACE` on demand to diagnose control plane behavior (e.g., API server errors, scheduling issues, controller reconciliation).
- **Incident response**: Quickly raise verbosity during an incident without cluster downtime or recreation.

### Requested Feature

Please add support for configuring the log level for EKS control plane components, ideally:

1. Dynamically (without cluster recreation) via `UpdateClusterConfig` or a dedicated API
2. Per-component granularity (e.g., `kube-apiserver`, `kube-scheduler`, `kube-controller-manager`)
3. Supported levels: `TRACE`, `DEBUG`, `INFO`, `WARN`, `ERROR`

### References

- EKS Control Plane Logging docs: https://docs.aws.amazon.com/eks/latest/userguide/control-plane-logs.html
- Kubernetes log verbosity: https://kubernetes.io/docs/concepts/cluster-administration/logging/

Start by reviewing the linked EKS Control Plane Logging and Kubernetes log verbosity documentation. The issue does not name repository files, tests, or an implementation entry point; first determine whether the requested configuration is exposed by an AWS API and define the scope for dynamic, per-component log levels. Done would require an agreed feature design and implementation path.

Written by the indexing model from the issue text.