Application Real-Time Monitoring Service:Obtain container service component metrics through custom collection tasks

Prerequisites

Before you begin, make sure that you have:

• Integrated Container Observability

Step 1: Copy the reference configuration

Copy the existing collection configuration to use as a template.

1. Log on to the ARMS console. In the left-side navigation pane, click Integration Management.

2. On the Access Management page, click the target access environment, and then click the Metric Collection tab.

3. In the left-side navigation pane, click Custom Collection, and then navigate to cs-default-monitor > _kube-state-metrics. Click View Configuration and copy the full configuration.

   

Step 2: Create a custom collection task

1. On the right side of the page, click Custom Collection, and then click Add.

2. On the Add Custom Collection Task page, set the following parameters.

   Parameter	Description	Example
   Task Name	Name of the custom task. Do not use uppercase letters.	ksm_demo
   Collection Configuration	Paste the configuration you copied in Step 1, then add a metric_relabel_configs block at the same level as relabel_configs to filter for the metrics you need. See the following sample configuration.	See below

   Sample configuration

   This configuration filters three kube-state-metrics values. Replace the metric names in the regex field with the ones your workloads require.

      scrape_configs:
      - job_name: ksm_job
        honor_timestamps: true
        scrape_interval: 30s
        scrape_timeout: 30s
        metrics_path: /metrics
        scheme: http
        follow_redirects: true
   
        ## Add this block to keep only the metrics you need
        metric_relabel_configs:
        - source_labels:
          - __name__
          regex: (kube_cronjob_next_schedule_time|kube_job_status_start_time|kube_job_spec_completions)
          action: keep
        ## Replace the metric names above with the ones required by your workloads
   
        relabel_configs:
        - action: labelmap
          regex: .*label_o11y_aliyun_dev_(.*)
          replacement: $1
        - source_labels:
          - __meta_kubernetes_service_label_app
          - __meta_kubernetes_service_label_k8s_app
          separator: ;
          regex: (kube-state-metrics);.*|.*;(kube-state-metrics)
          replacement: $1
          action: keep
        - source_labels:
          - __meta_kubernetes_pod_container_port_number
          separator: ;
          regex: 8080
          replacement: $1
          action: keep
        - source_labels:
          - __address__
          - __meta_kubernetes_pod_container_port_number
          separator: ;
          regex: ([^:]+)(?::\d+)?;(\d+)
          target_label: __address__
          replacement: $1:$2
          action: replace
        kubernetes_sd_configs:
        - api_server: https://kubernetes.default:8443
          role: endpoints
          authorization:
            type: Bearer
            credentials_file: /var/run/secrets/kubernetes.io/serviceaccount/token
          tls_config:
            insecure_skip_verify: true
          follow_redirects: true
          namespaces:
            names:
            - cs-ask-c8034dddd2cf646318b63faf1********

   The key addition is the metric_relabel_configs block. It uses a keep action with a regex that matches only the specified metric names in source_labels: [__name__]. All other metrics scraped by this job are dropped.

3. Click Check And Create.

Step 3: Verify the collection task

1. Click the Self-monitoring tab and confirm that the State of your new task shows UP.

   

2. Click the Metric Explorer tab and query the metrics you configured. Use the job name from your custom task (for example, ksm_job) as the filter value.

   Note

   The filter value corresponds to the job name specified in your custom collection task.