Add InstanceType to GPU metrics

[movence]

Description:
Add InstanceType attribute to GPU metrics

Testing:
Tested on a test cluster. Example EMF entry:

{
    "CloudWatchMetrics": [
        {
            "Namespace": "ContainerInsights",
            "Dimensions": [
                [
                    "ClusterName"
                ],
                [
                    "ClusterName",
                    "InstanceId",
                    "NodeName"
                ],
                [
                    "ClusterName",
                    "GpuDevice",
                    "InstanceId",
                    "InstanceType",
                    "NodeName"
                ]
            ],
            "Metrics": [
                {
                    "Name": "node_gpu_temperature",
                    "Unit": "None"
                },
                {
                    "Name": "node_gpu_power_draw",
                    "Unit": "None"
                },
                {
                    "Name": "node_gpu_utilization",
                    "Unit": "Percent"
                },
                {
                    "Name": "node_gpu_memory_utilization",
                    "Unit": "Percent"
                },
                {
                    "Name": "node_gpu_memory_used",
                    "Unit": "Bytes"
                },
                {
                    "Name": "node_gpu_memory_total",
                    "Unit": "Bytes"
                }
            ]
        }
    ],
    "ClusterName": "mixed-test",
    "GpuDevice": "nvidia0",
    "InstanceId": "i-xxxxxxxxxx",
    "InstanceType": "g4dn.xlarge",
    "NodeName": "ip-x-x-x-x.us-west-2.compute.internal",
    "OTelLib": "otelcol/prometheusreceiver",
    "Timestamp": "1710445148938",
    "Type": "NodeGPU",
    "Version": "0",
    "http.scheme": "https",
    "k8s.namespace.name": "amazon-cloudwatch",
    "kubernetes": {
        "host": "ip-x-x-x-x.us-west-2.compute.internal"
    },
    "net.host.name": "dcgm-exporter-service.amazon-cloudwatch.svc",
    "net.host.port": "9400",
    "service.instance.id": "dcgm-exporter-service.amazon-cloudwatch.svc:9400",
    "service.name": "containerInsightsDCGMExporterScraper",
    "node_gpu_memory_total": 16106127360,
    "node_gpu_memory_used": 198180864,
    "node_gpu_memory_utilization": 0.0007999999999999999,
    "node_gpu_power_draw": 33.199,
    "node_gpu_temperature": 31,
    "node_gpu_utilization": 9
}

[jefchien]

You'll need to fix the unit tests to accept the interface change.