DevOpsProdigy KubeGraf Node's Dashboard is broken

[dmitrievav]

kubelet_running_pod_count, container_cpu_usage_seconds_total and container_memory_usage_bytes have no label node

Example:

kubelet_running_pod_count{beta_kubernetes_io_arch="amd64",beta_kubernetes_io_instance_type="m5a.large",beta_kubernetes_io_os="linux",failure_domain_beta_kubernetes_io_region="eu-central-1",failure_domain_beta_kubernetes_io_zone="eu-central-1a",instance="ip-10-102-36-19.eu-central-1.compute.internal",job="kubernetes-nodes",kops_k8s_io_instancegroup="nodes",kubernetes_io_hostname="ip-10-102-36-19.eu-central-1.compute.internal",kubernetes_io_role="node"}

[SergeiSporyshev]

Hi! Thank you for your PR
I think this issue (#9) will help you

[dmitrievav]

Adding to my prometheus.yml following rules solved the issue

- job_name: 'kubernetes-nodes-cadvisor'
  ...
  relabel_configs:
    ...
    # add node label for DevOpsProdigy KubeGraf compatibility
    - action: labelmap
      regex: __meta_kubernetes_node_label_kubernetes_io_hostname
      replacement: node

- job_name: 'kubernetes-nodes'
  ...
  relabel_configs:
    ...
    # add node label for DevOpsProdigy KubeGraf compatibility
    - action: labelmap
      regex: __meta_kubernetes_node_label_kubernetes_io_hostname
      replacement: node

And here is Prometheus documentation: https://prometheus.io/docs/prometheus/latest/configuration/configuration/#relabel_config