Config validate fails with resourcedetection k8snode

[martinohansen]

Component(s)

processor/resourcedetection

What happened?

Description

The validate command fails if not connected to Kubernetes and resourcedetection contains detector k8snode.

Steps to Reproduce

Run validate with this processor in your config

resourcedetection/k8s:
  detectors: ["k8snode"]
  override: true

Expected Result

nil

Actual Result

2025/06/10 11:40:34 collector server run finished with error: failed to build pipelines: failed to create "resourcedetection/k8s" processor, in pipeline "metrics": failed creating detector type "k8snode": failed creating k8snode detector: failed to create K8s API client: unable to load k8s config, KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT must be defined

Collector version

0.126.0

[github-actions]

Pinging code owners:

• processor/resourcedetection: @Aneurysm9 @dashpole

See Adding Labels via Comments if you do not have permissions to add labels yourself.

[dashpole]

It looks like it is expecting KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT to be defined, which are always defined when the collector is running in a pod. Where are you trying to run this from?

[martinohansen]

CI to validate that the config is valid before deploying it. I've only seen this processor fail the config validation, k8sattributes for examples does break/fail it.

[dashpole]

Looks like a common problem for a few of the metadata providers. E.g. kubeadm does the same thing:

opentelemetry-collector-contrib/internal/metadataproviders/kubeadm/metadata.go

Line 36 in d50708b

k8sAPIClient, err := k8sconfig.MakeClient(apiConf)

. The best solution might be to move creation of the resource provider to the resourcedetection processor's Start() function:

opentelemetry-collector-contrib/processor/resourcedetectionprocessor/resourcedetection_processor.go

Line 30 in d50708b

func (rdp *resourceDetectionProcessor) Start(ctx context.Context, host component.Host) error {