This topic describeshow to connect your additional O&M monitoring component to an edge cluster by using the edge O&M tunnel. This allows you to enhance the capability of monitoring applications on edge nodes.

Background information

  • In a managed edge cluster, metrics-server is a default component that can obtain the monitoring data of containers by using kubelet. The data includes CPU, memory, and other information.
  • The configurations in this topic are applicable to the following two scenarios: When you migrate services to the cloud, you must also migrate the O&M monitoring system to ensure a seamless service migration. When you want to collect more monitoring data, such as that of service applications, you must connect self-developed or open source monitoring components to the cluster.
Notice

To use the O&M tunnel of an edge cluster to integrate an additional O&M monitoring component, make sure that the following requirements are met (not applicable to other monitoring solutions):

  • Deploy the O&M monitoring component and the O&M tunnel component edge-tunnel-server on the same node in the cloud.
  • The O&M monitoring component obtains the monitoring data of edge services by using HTTP pull requests.

Deployment model

The following figure shows how to connect an additional O&M monitoring component to the cluster.Deployment model

Configure an additional O&M monitoring component

You must configure the additional O&M monitoring component in advance based on the access mode that the component uses to monitor objects. Configure the additional component based on the following instructions.
  • Use Node IP address:Port to access monitoring data

    The monitoring component uses http://Node IP address:Port/{path} as an endpoint to obtain monitoring data.

    • Deploy the applications on pods that share networks with the hosts.
    • By default, ports 10250 and 10255 are supported.
    • For other ports, you must configure the edge-tunnel-server-cfg ConfigMap to connect the additional O&M monitoring component. The following example describes how to configure parameters.
      #Example: The monitoring component uses http://{Node IP address}:8080/metrics as an endpoint to obtain the monitoring data of applications. Configure the following parameters.
      cat <<EOF | kubectl apply -f
      apiVersion: v1
      data:
        dnat-ports-pair: '8080=10264' #Access port 1=10264, Access port 2=10264
      kind: ConfigMap
      metadata:
        name: edge-tunnel-server-cfg
        namespace: kube-system
      EOF
  • Use Pod IP address:Port to obtain monitoring data.

    The monitoring component uses http://Pod IP address:Port/{path} as an endpoint to obtain monitoring data.

    Currently, this mode is unavailable.