Docs

Volcengine Kubernetes Engine

Monitoring

Monitoring VCI with VMP

Basic VCI monitoring can be implemented based on VMP. This topic describes how to configure collection rules for VCI monitoring metrics and how to query metrics.

Prerequisites

Cloud-native observability has been enabled for the VCI business cluster based on VCI. For details, see Enabling cloud-native observability.

The prometheus-agent add-on in the VCI business cluster has been upgraded to version v2.2.0 or later. For details, see Add-on release records.

Configuring collection rules

After you have enabled cloud-native observability for the VCI business cluster and installed the corresponding add-ons, you need to configure the relevant collection rules to collect metrics properly. The following procedure describes how to configure collection rules in VCI scenarios. For details, see Observability and AI resources observability.

In the left navigation bar, click Cluster.

On the cluster list page, click the target cluster.

In the left navigation bar of the cluster management page, click Observability settings.

Enable the observability add-ons as needed and edit the VCI metric configuration. For detailed steps and descriptions, see VKE observability and AI resources observability.

tip

In VCI scenarios, you mainly use the Container service and AI resources observability add-ons.

Non-GPU metric settings (such as CPU, memory, and network) can be edited in the VKE add-ons.

GPU-related metric settings can be edited in the AI resources observability add-ons.

For detailed monitoring metrics supported by VCI, see Getting VCI metrics through virtual nodes and VCI GPU monitoring metrics.

Querying metrics

After configuring the monitoring collection rules, you can query VCI metrics in the VMP console. For details, see Querying metrics.

In the top navigation bar, choose the target region.

In the left navigation bar, click Explore to go to the Explore page.

In the upper-right corner of the page, specify the workspace, time range, and data refresh method for the metric query.

Use PromQL query statements to query VCI metrics.

Query VCI metrics via virtual nodes by using $metric_name{node="$virtual_node_name"}, for example, container_cpu_usage_seconds_total{node="vci-node1-cn-beijing-a"}.

Query VCI GPU metrics by using $metric_name{pod="$VCI_instance_name"}, for example, DCGM_FI_DEV_DEC_UTIL{pod="vci-ini2-d75d77bc5-82lch"}.

Click Query to query the metrics and display the metrics dashboard.

Last updated: 2026.07.24 16:18:04