While doing kubectl cluster-info dump , I see alot of:
2018/10/18 14:47:47 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:48:17 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:48:47 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:49:17 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:49:47 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:50:17 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
2018/10/18 14:50:47 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
Maybe this is a bug that will be fixed in new version ( heapster is deprecated anyway in new versions) , but is there anyway to disable these checks to avoid these noisy messges.
You can find Heapster deprecation timeline here.
I found that in Kubernetes cluster 1.10 version kubernetes-dashboard Pod produces such kind of error messages:
kubectl --namespace=kube-system log <kubernetes-dashboard-Pod>
2018/10/22 13:04:36 Metric client health check failed: the server could not find the requested resource (get services heapster). Retrying in 30 seconds.
It seems that kubernetes-dashboard still requires Heapster service for metrics and graph purposes.