What do metadata-agent health errors in GKE logs mean?

1/3/2019

I setup a standard GKE cluster including istio. In the logs I find errors among which e.g.:

{
 insertId:  "xuxoovg5olythg"  
 logName:  "projects/projectname/logs/stderr"  
 metadata: {…}  
 receiveTimestamp:  "2019-01-03T11:39:10.996283280Z"  
 resource: {
  labels: {
   cluster_name:  "standard-cluster-1"    
   container_name:  "metadata-agent"    
   location:  "europe-west4-a"    
   namespace_name:  "kube-system"    
   pod_name:  "metadata-agent-859mq"    
   project_id:  "myprojectname"    
  }
  type:  "k8s_container"   
 }
 severity:  "ERROR"  
 textPayload:  "W0103 11:39:06 7fc26c8bf700 api_server.cc:183 /healthz returning 500; unhealthy components: Pods
"  
 timestamp:  "2019-01-03T11:39:06Z"  
}

Is this a problem? What to do with this?

-- musicformellons
google-cloud-platform
google-kubernetes-engine
istio
kubernetes

1 Answer

1/23/2019

Currently it is said to be working as intended, as a short-term solution. Possibly will be fixed in the future.

Source: https://github.com/Stackdriver/metadata-agent/issues/183#issuecomment-426406327

-- Jakub Gocławski
Source: StackOverflow