[processor/resourcedetection] instance hostname missing on GKE with new detector

[mx-psi]

Describe the bug

Before #10347, someone using the gke and gce detectors would have the instance hostname on GKE. After this change, the new gcp detector would not add this attribute. This is a regression on v0.55.0, and also makes it difficult for people to transition to the new detector.

Datadog needs this attribute in order to correctly correlate telemetry reported by the Collector with telemetry reported by the Datadog GCP integration, which uses this as part of the hostname.

Steps to reproduce

Add the gke and gce detectors to your Collector on a GKE cluster.

What did you expect to see?

The host.name attribute would be present as it happened on v0.54.0 and below

What did you see instead?

The host.name attribute is missing.

What version did you use?

v0.55.0

What config did you use?

receivers:
  hostmetrics:
    scrapers:
      load:
    collection_interval: 10s

processors:
  resourcedetection:
    detectors: [gke, gce]

exporters:
  logging:
    loglevel: debug

service:
  pipelines:
    metrics:
      receivers: [hostmetrics]
      processors: [resourcedetection]
      exporters: [logging]

Environment

v0.55.0 Docker image