Skip to content

Monitoring Edge Resource

This document explains how to enable monitoring capabilities in edge units to view the status and resource usage of edge nodes and workloads in real time.

Edge monitoring relies on the insight-agent component and the edge monitoring component. The following sections describe how to install the proper components.

Note

Install the insight-agent component first, as the edge monitoring component depends on Insight-related plugins.

Check if Monitoring Components are Installed

Select Cloud Edge Collaboration from the left navigation bar to enter the edge unit page. Check the Insight and edge component information in the edge unit entries:

  • If a version number is displayed, the monitoring component is installed.
  • If Install Now is displayed, the monitoring component is not installed. Click the Install Now button to navigate to the proper Helm app details page.

Install insight-agent

The insight-agent is a plugin for cluster observation data collection, supporting unified monitoring of metrics, traces, and logging. The installation process can be referenced from Online Installation of insight-agent.

To prevent the insight DaemonSet component from being scheduled to edge nodes, add the following affinity settings to each DaemonSet component under the insight-system namespace of the Global Service Cluster.

nodeAffinity:
  requiredDuringSchedulingIgnoredDuringExecution:
    nodeSelectorTerms:
      - matchExpressions:
        - key: node-role.kubernetes.io/edge
          operator: DoesNotExist

Install Edge Monitoring Component

  1. Select Container Management -> Clusters from the left navigation bar to enter the clusters page. Click the cluster name to enter the cluster details page.

  2. Select Helm Apps -> Helm Charts from the left menu. Find the insight-edge-agent plugin under the addon repository and click to enter the chart details page.

  3. In the top right corner of the page, select the insight-edge-agent version and click the Install button to enter the insight-edge-agent installation page.

  4. Fill in the addresses of the proper data storage components in the global management cluster and the edge worker cluster in the form. After confirming that the information is correct, click OK .

    To achieve unified storage and querying of observation data across multiple clusters, the edge cluster needs to report the collected observation data to the global management cluster for unified storage.

    • Obtain the Prometheus service address and port in the Edge Worker Cluster . Currently, only NodePort is supported, so the address corresponds to that of the worker cluster control node.

      Enter the following command in the insight-system namespace to find the service port mapped to port 9090:

      kubectl get svc -n insight-system | grep prometheus
      

      Note

      If the default access method for the Prometheus service is ClusterIP, please change it to NodePort.

    • Obtain the ElasticSearch service address and port in the Global Service Cluster. Currently, only NodePort is supported, so the address corresponds to that of the cluster controller node.

      Enter the following command in the insight-system or mcamel-system namespace to find the service port mapped to port 9200:

      kubectl get service -n mcamel-system | grep es
      

      Note

      Other parameters do not need to be modified by default.

  5. The system will automatically return to the Helm app list. When the status of insight-edge-agent changes from Pending to Deployed , and all component statuses are Running , the installation is successful.

View Edge Monitoring Data

After the edge components are successfully installed and deployed, wait for a while, and then you can enter the Insight module from the left navigation bar to view the data of edge resources.

Comments