Docs

Volcengine Kubernetes Engine

Basic observation

Overview of basic observability

VKE supports cloud-native observability. This enables you to collect, monitor, and trigger alerts on metrics for clusters, basic resources, and applications.

Basic principles

Basic observability is built on Volcengine Managed Service for Prometheus (VMP), which is compatible with open-source Prometheus APIs and the community Exporter ecosystem. This enables rapid monitoring of cluster infrastructure resources and user workloads.

Compared with a self-managed Prometheus environment, you do not manually need to download or configure Prometheus, or worry about the storage cost of large volumes of data. Basic observability provides a setup-free, maintenance-free integrated Prometheus environment, reducing development and O&M costs.

Main features

The main features of cloud-native basic observability are shown in the table below.

Category	Feature	Basic observability
Installation and deployment	Data collection	The prometheus-agent add-on.
		High availability	Unmanaged add-ons, which support sharding and automatic scaling.
		Data storage	Data storage is implemented based on VMP. Workspaces with different specifications are supported. For details, see Workspace specifications.
Monitoring	Cluster monitoring	Cluster resource monitoring and custom application monitoring. Valid values: Basic resources, including clusters, nodes, workloads, pods, containers, volumes, and more. Control plane add-ons, including kube-apiserver, kube-scheduler, etcd, and more. Other resources, including GPU, P2P acceleration, Nydus, RDMA, and more.
		Custom monitoring	Custom cloud-native application monitoring is supported. Application metrics can be exposed by using the Prometheus SDK to enable application monitoring.
		Self-built Exporter monitoring	Various types of Exporters provided by the open-source Prometheus community can be used to collect metrics from different services. Examples of these services are databases, hardware, message middleware, storage systems, and HTTP servers. Custom Exporters can be created based on the Client Library available in the open-source Prometheus community. The Prometheus community provides support for Go, Java, Python, Ruby, and other languages.
Metric viewing	Built-in dashboards	Supported, with a wide variety.
		Custom dashboards	Supported. You can create one based on open-source Grafana.
		Metric query	Metrics can be queried by using Explore.
Alert center	Metric alerting	Supported
		Alert levels	Supported
		Alerting templates	Supported
		Alert silencing	Supported
		Alert notification methods	Email, Webhook, phone, Feishu, DingTalk, WeCom, and more.

Last updated: 2026.07.24 10:53:28