Skip to content

insight-agent Component Status Explanation

In AI platform, Insight acts as a multi-cluster observability product. To achieve unified data collection across multiple clusters, users need to install the Helm App insight-agent (installed by default in the insight-system namespace). Refer to How to Install insight-agent .

Status Explanation

In the "Observability" -> "Collection Management" section, you can view the installation status of insight-agent in each cluster.

  • Not Installed : insight-agent is not installed in the insight-system namespace of the cluster.
  • Running : insight-agent is successfully installed in the cluster, and all deployed components are running.
  • Error : If insight-agent is in this state, it indicates that the helm deployment failed or there are components deployed that are not in a running state.

You can troubleshoot using the following steps:

  1. Run the following command. If the status is deployed , proceed to the next step. If it is failed , it is recommended to uninstall and reinstall it from Container Management -> Helm Apps as it may affect application upgrades:

    helm list -n insight-system

  2. Run the following command or check the status of the deployed components in Insight -> Data Collection . If there are Pods not in the Running state, restart the containers in an abnormal state.

    kubectl get pods -n insight-system

Additional Notes

  1. The resource consumption of the Prometheus metric collection component in insight-agent is directly proportional to the number of Pods running in the cluster. Please adjust the resources for Prometheus according to the cluster size. Refer to Prometheus Resource Planning.

  2. The storage capacity of the vmstorage metric storage component in the global service cluster is directly proportional to the total number of Pods in the clusters.