Docs

Volcengine Managed Service for Prometheus

High-availability collection solution

Deploying VM Agent

To use a high-availability collection solution, you need to first deploy the VM Agent collector in the target cluster. This topic describes the basic concepts, deployment procedure, and parameter configuration of the VM Agent collector in detail.

warning

VM Agent is a native version of VictoriaMetrics and does not support AK/SK authentication of Volcano Engine.

If you use a VKE cluster, we recommend using the add-on management feature in clusters to install collectors and add-ons. For more details, see VKE integration.

Overview

High availability description

The VM Agent collector comes with an operator, which achieves sharding and data collection for the collector and the kube-state-metrics add-on, achieving high performance and availability.

Agent description

When you install the VM Agent collector in the cluster, the vm-operator, kube-state-metrics, and node-exporter add-ons are also installed. The following table describes the name, type, and feature of each agent.

Agent Name	Workload Type	Namespace	Description
vmagent	StatefulSet	volcano-metrics	Collects metrics exposed by kubelet, kubelet-cadvisor, kube-state-metrics, and node-exporter in the cluster.
vm-operator	Deployment	volcano-metrics	The controlling agent of VM Agent.
kube-state-metrics	StatefulSet	volcano-metrics	Provides the metric data of Kubernetes resources.
node-exporter	DaemonSet	volcano-metrics	Provides the metric data of node resources.

VM Agent serves as a data collection agent of Volcengine Managed Service for Prometheus (VMP). It scrapes monitoring data of monitoring targets and sends the data to VMP for storage. By default, the monitoring data includes the data of node-exporter, kubelet metrics/cadvisor, kubelet metrics, kube-state-metrics, and the agent itself.

Default specifications

The following table shows the default instance specifications of the add-ons related to the VM Agent collector.

Agent Name	Default CPU Request	Default Memory Request	Default CPU Upper Limit	Default Memory Upper Limit
vmagent	2 core	2 GiB	4 core	4 GiB
kube-state-metrics	0.2 Core	512 MiB	0.8 core	2 GiB
node-exporter	0.05 core	128 MiB	0.1 core	256 MiB

Prerequisites

A self-managed Kubernetes cluster is created, and a virtual private cloud (VPC) is connected to an Internet data center (IDC) through methods such as Cloud Enterprise Network (CEN).

A VMP workspace has been created. For details, see Creating workspaces.

When you use the script to install VM Agent, you need to use wget, echo, curl, and sed in your local environment (Linux/Mac). Ensure that these command-line tools are installed on your system.

warning

The VM Agent collector cannot be used together with the prometheus-agent add-on in the VKE cluster. Before you deploy the VM Agent collector, make sure that the prometheus-agent add-on is not installed in your cluster. For more details, refer to VKE integration.

Deploy the instance (default configuration)

You can use the deployment script to quickly deploy the VM Agent collector. You can perform the following steps to deploy the default configurations:

Download the script in your local environment. For more details, refer to Obtaining a script.

Use kubectl to connect to the Kubernetes cluster.

Run the following command to deploy the VM Agent collector:

tip

Modify the region, remote-write-url, namespace, basic-auth-username, and basic-auth-password parameters in the preceding configuration based on your business requirements.

Deploy the instance (custom configuration)

If the default specification of the collector cannot meet your business requirements, you can use predefined parameters in the deployment script to modify the instance configuration (such as the CPU and memory capacity of the collector) before deployment. Procedure:

Download the script in your local environment. For more details, refer to Obtaining a script.

Use kubectl to connect to the Kubernetes cluster.

Run the following command to modify parameters and deploy the VM Agent collector:

Deploy the instance (custom parameters)

If the predefined parameters in the deployment script cannot meet your business requirements, you can perform the following steps to specify custom parameters:

Download the script in your local environment. For more details, refer to Obtaining a script.

Use kubectl to connect to the Kubernetes cluster.

Run the following commands to specify your custom parameters and generate the install.yaml configuration file:

Open the install.yaml file, specify relevant parameters, and save the file.

Run the following command to deploy the related agents:

Check the result

After VM Agent is deployed, log in to the cluster and you can view the workloads of vm-operator in the list of Deployments. In the list of StatefulSets, you can view the workloads of VM Agent and kube-state-metrics.

tip

To ensure that the collector can collect data as expected, you need to make sure that the collector runs as expected and the metrics API or the exporter of the collection target is functioning correctly. You must complete the related configuration and deployment, and ensure they are configured correctly.

What to do next

After you deploy the collector, in addition to the default collection rules, you can also use ServiceMonitor or PodMonitor to configure collection rules for custom targets. For more details, refer to Service discovery.

Use the --help parameter

You can view all the predefined fields and their meanings in the installation script by using the --help parameter.

The following section describes the meanings and examples of different fields in the installation script.

Parameters	Description
--agent	Select the collector to deploy. The value is fixed as vmagent.
--remote-write-url	The remote write URL of the VMP workspace. For details, see Obtaining workspace addresses.
--basic-auth-username	The username for authentication in Basic Auth mode of the VMP workspace.
--basic-auth-password	The password for authentication in Basic Auth mode of the VMP workspace.
--region	The region of the collector cluster, such as cn-beijing. For more details, refer to Regions and Availability Zones.
--exec-script-env	The script execution environment, including: local: the local environment. It indicates that you execute the script in your local environment. public-cloud-vpc: the VPC. It indicates that you execute the script in the VPC.
--rendering-only	You can only view the installation list of add-ons, such as VM Agent and vm-operator, and do not install the add-ons.
--namespace	The namespace of the collector. If you do not specify this parameter, the collector will be deployed in the volcano-metrics namespace by default.
--storage-class	The PersistentVolumeClaim (PVC) storage type.
--kube-state-metrics-shards	The number of shards for the kube-state-metrics agent.
--kube-state-metrics-resources-requests-cpu	The CPU request of kube-state-metrics. The value must be an integer. Unit: m. For example, 100m indicates a CPU request of 100 millicores.
--kube-state-metrics-resources-requests-memory	The memory request of kube-state-metrics. The value must be an integer. Unit: Mi. For example, 200Mi indicates a memory request of 200 MB.
--kube-state-metrics-resources-limits-cpu	The CPU upper limit of kube-state-metrics. The value must be an integer. Unit: m. For example, 100m indicates a maximum value of 100 millicores.
--kube-state-metrics-resources-limits-memory	The memory upper limit of the kube-state-metrics collector. The value must be an integer. Unit: Mi. For example, 200Mi indicates a maximum value of 200 MB.
--agent-scrape-interval	The data collection interval of the collector. The value must be an integer in seconds. Example: 15 seconds.
--vmpagent-name	The name of the VMP Agent collector.
--vmpagent-replicas	The number of replicas for the VMP Agent collector.
--vmpagent-pvc-capacity	The storage capacity of the VMP Agent collector.
--vmpagent-resources-requests-cpu	The CPU request of the VMP Agent collector. The value must be an integer. Unit: m. For example, 100m indicates a CPU request of 100 millicores.
--vmpagent-resources-requests-memory	The memory request of the VMP Agent collector. The value must be an integer. Unit: Mi. For example, 200Mi indicates a memory request of 200 MB.
--vmpagent-resources-limits-cpu	The CPU upper limit of the VMP Agent collector. The value must be an integer. Unit: m. For example, 100m indicates a maximum value of 100 millicores.
--vmpagent-resources-limits-memory	The memory upper limit of the VMP Agent collector. The value must be an integer. Unit: Mi. For example, 200Mi indicates a maximum value of 200 MB.
--vmagent-replicas	The number of replicas for the VM Agent collector.
--vmagent-shards	The number of shards for the VM Agent collector.
--vmagent-pvc-capacity	The storage capacity of the VM Agent collector.
--vmagent-resources-requests-cpu	The CPU request of the VM Agent collector. The value must be an integer. Unit: m. For example, 100m indicates a CPU request of 100 millicores.
--vmagent-resources-requests-memory	The memory request of the VM Agent collector. The value must be an integer. Unit: Mi. For example, 200Mi indicates a memory request of 200 MB.
--vmagent-resources-limits-cpu	The CPU upper limit of the VM Agent collector. The value must be an integer. Unit: m. For example, 100m indicates a maximum value of 100 millicores.
--vmagent-resources-limits-memory	The memory upper limit of the VM Agent collector. The value must be an integer. Unit: Mi. For example, 200Mi indicates a maximum value of 200 MB.
--disable-cadvisor	Disables the rule that allows VMP Agent to collect the metrics of cAdvisor.
--cadvisor-only	Only enables the rule that allows VMP Agent to collect the metrics of cAdvisor.
--disable-install-node-exporter	Disables the rule that allows VMP Agent to collect the metrics of node-exporter.

Last updated: 2026.10.10 10:16:50