[BUG] CLUSTER | ERROR | timeout or abort while handling URI /apis/external.metrics.k8s.io/v1beta1

[seifrajhi]

Agent Environment

Datadong agent image: public.ecr.aws/datadog/cluster-agent:7.49.0
chart version: 3.45.0

Describe what happened:

We receive many error messages :

 CLUSTER | ERROR | (apimachinery@v0.27.6/pkg/util/runtime/runtime.go:109 in HandleError) | post-timeout activity - time-elapsed: 2.288897ms, GET "/apis/external.metrics.k8s.io/v1beta1" result: <nil>

CLUSTER | ERROR | (apimachinery@v0.27.6/pkg/util/runtime/runtime.go:109 in HandleError) | timeout or abort while handling: method=GET URI="/apis/external.metrics.k8s.io/v1beta1" 

CLUSTER | ERROR | (apimachinery@v0.27.6/pkg/util/runtime/runtime.go:109 in HandleError) | apiserver received an error that is not an metav1.Status: &errors.errorString{s:"http2: stream closed"}: http2: stream closed

agent status output:

 agent status
 ...
 ...
Custom Metrics Server
=====================

  Data sources
  ------------
  URL: https://api.datadoghq.eu

  
  External metrics provider uses DatadogMetric - Check status directly from Kubernetes with: `kubectl get datadogmetric`

Describe what you expected:

NA

Steps to reproduce the issue:

Run datadog helm chart datadog-3.45.0 in AWS EKS cluster v1.25