Elastic Container Instance:Monitor disks with Prometheus

Prerequisites

• You have created an ACK Serverless cluster and connected the cluster to ARMS Prometheus. For more information, see Connect to Prometheus Monitoring.

• You have created a dedicated (non-shared) Grafana workspace, completed data migration, and integrated Alibaba Cloud services. For more information, see Create a Grafana workspace and Integrate Alibaba Cloud services.

  Important

  • By default, ARMS Prometheus uses a shared Grafana workspace, which does not support custom dashboards.

  • The steps may differ for other Grafana versions.

Procedure

1. Log on to the ARMS console.

2. Open the Grafana page.

   1. In the left navigation pane, choose Managed Service for Grafana > Workspace Management.

   2. Find your workspace and click its URL to open the Grafana logon page.

3. Enter your credentials to log on to Grafana.

   Use the admin credentials you configured when creating the workspace, or use a different account.

4. In the upper-left corner of the Grafana page, click the icon, and select Dashboards.

5. On the Dashboards page, click New in the upper-right corner and select Import.

6. Click Upload dashboard JSON file and select your file.

   The JSON file contains the dashboard configuration. Sample file: ECI_Pod_Disk.

7. Select a folder for the dashboard and choose your cluster as the data source.

8. Click Import.

   After the dashboard is imported, it displays disk monitoring data for your ECI pods.

   

9. (Optional) To filter monitoring data by Deployment or StatefulSet, you can either copy the relevant panel to a different dashboard or modify the current dashboard's settings directly.

   To add filter variables by modifying the dashboard's settings, follow these steps:

   1. In the disk monitoring dashboard in Grafana, click the icon in the upper-right corner.

   2. On the Settings page, choose Variables in the left navigation pane.

   3. Click New variable, configure the settings in the General and Query Options sections, and then click Apply.

      The following table provides the queries required to create deployment and statefulset variables.

      Type	Query type	Label *	Metric
      deployment	Label_values	deployment	kube_deployment_created{namespace=~"$namespace"}
      statefulset	Label_values	statefulset	kube_statefulset_created{namespace=~"$namespace"}

   4. Click the pod variable, modify the Query, and then click Apply.

      Modify the query for the pod variable to filter based on the variable you just created. The following table provides the required queries.

      Type	Query type	Query
      deployment	Query_result	kube_pod_info{namespace=~'$namespace',pod=~"$deployment.*"} >= 1
      statefulset	Query_result	kube_pod_info{namespace=~'$namespace',pod=~"$statefulset.*"} >= 1

   5. Move the new deployment (or statefulset) variable so it appears above the pod variable.

      Note

      You can click and hold the icon to drag the variable to a new position.

      The following code block shows the final order and definitions for the variables.

      Variable        Definition
      namespace       label_values(kube_pod_info, namespace)
      deployment      label_values(kube_deployment_created{namespace=~"$namespace"},deployment)
      pod             query_result(kube_pod_info{namespace=~'$namespace',pod=~"$deployment.*"} >= 1)

   6. In the upper-right corner, click Save dashboard, enter a description of your changes, and click Save.

   7. Return to the dashboard to verify that the filter works.

      The following figure shows the Deployment-level filter: