You can install various add-ons in clusters, including network, storage, monitoring, DNS, security, image, and GPU add-ons, to meet your different business requirements. You can deploy, upgrade, or uninstall add-ons based on your business requirements.

Viewing add-ons

You can view the add-ons provided by Volcengine Kubernetes Engine (VKE) based on the following descriptions. To view the release records of each add-on, see Add-on release records.

Log in to the VKE console.

In the left navigation bar, click Cluster. On the page that appears, find the cluster for which you want to view add-ons and click the name of the cluster.

In the left navigation bar of the cluster management page, click Addons. On the page that appears, you can view various add-ons provided by VKE.

On the Addons page of a VPC-CNI network model cluster, some add-ons have the VCI label. You can deploy these add-ons by using the following methods:

Elastic Compute Service (ECS) deployment: deploys the add-ons on worker nodes (ECS nodes).

Volcengine Container Instance (VCI) deployment: deploys the add-ons on VCI.

Usage limits

You cannot install add-ons in a Flannel network model cluster when the cluster has no ECS nodes.

You can install add-ons with the VCI label only on VCI in a VPC-CNI network model cluster. You cannot install other add-ons.

Add-ons and plug-ins that are deployed as DaemonSets are not supported on VCI. For more information, see VCI limits.

Add-on introduction

The following table lists various add-ons provided by VKE.

Network

Add-on name	Add-on description	Deployment method	Installation recommendations
alb-ingress-controller	The alb-ingress-controller functions as an ALB Ingress controller add-on. It allows clusters to create and manage ALB instances. This achieves traffic management and application release based on ALB instances.	Users do not need to be concerned with this.	Required
ingress-nginx	The ingress-nginx add-on is an Ingress controller that uses Nginx as a reverse proxy and load balancer. warning NGINX Ingress support will soon be discontinued. For details, see [Product announcement] NGINX Ingress EOM.	ECS VCI	Optional
apig-controller	The apig-controller functions as an APIG Ingress controller add-on, which allows clusters to create and manage APIG instances for traffic management and application release.	ECS VCI	Optional
vpc-cni	The vpc-cni add-on is a network add-on of clusters, which supports the VPC-CNI network model. tip This add-on is required in clusters that use the VPC-CNI network model. The add-on is not displayed if the cluster uses the Flannel network model.	Users do not need to be concerned with this.	Required
flannel	The flannel is a cluster network add-on that supports the Flannel network model. tip This add-on is required in clusters that use the Flannel network model. The add-on is not displayed if the cluster uses the VPC-CNI network model.	Users do not need to be concerned with this.	Required
vpc-cni-controlplane	[Beta - apply for trial] The vpc-cni-controlplane add-on is the VPC-CNI control plane add-on of clusters, which provides control plane management capability for configuring security groups for pods. warning This add-on is required when you create a cluster with the Trunk ENI feature enabled. You cannot modify the configurations of the add-on after installation.	Users do not need to be concerned with this.	Optional
rdma-device-plugin	[Public beta] The rdma-device-plugin add-on is a Remote Direct Memory Access (RDMA) device add-on, which manages RDMA network interface cards (NICs) and enables pod containers to use RDMA devices.	ECS	Optional

Storage

Add-on name	Add-on description	Deployment method	Installation recommendations
csi-fsx	The csi-fsx is a general-purpose FSX storage add-on. It enables container clusters to use file systems that support the FSX client through the standard Container Storage Interface (CSI).	ECS	Optional
fsxset-controller	‌FSX client fine-grained scheduling and O&M add-on. It is primarily used to manage the ‌lifecycle of FSX clients‌ (including creation, deletion, and hot upgrades).	ECS	Optional
csi-nas	The csi-nas add-on enables container clusters to use the NAS service through the standard CSI.	ECS VCI	Recommended
csi-ebs	The csi-ebs add-on enables container clusters to use the EBS service through the standard CSI.	ECS VCI	Recommended
csi-cloudfs	[Beta - apply for trial] The csi-cloudfs add-on is an add-on of CloudFS, which allows container clusters to use CloudFS via the standard CSI.	ECS	Optional
csi-vepfs	[Beta - apply for trial] The csi-vepfs is an add-on of vePFS, which allows container clusters to use the vePFS service through the standard CSI. tip After the add-on is uninstalled, running pods can continue to access vePFS until they stop running.	ECS	Optional
csi-local	[Beta·apply for trial] The csi-local add-on is a local storage service add-on. It allows container clusters to use local storage services by using the standard CSI API.	ECS	Optional
snapshot-controller	A storage snapshot add-on. It supports using storage snapshots in Volcengine Kubernetes Engine. warning The snapshot-controller add-on requires the csi-ebs add-on. You need to install both when setting up snapshot-controller. Uninstalling csi-ebs will automatically remove snapshot-controller.	ECS VCI	Recommended

Monitoring

Add-on name	Add-on description	Deployment method	Installation recommendations
metrics-server	The metrics-server add-on is a core cluster monitoring add-on that aggregates key monitoring metrics for resources across the cluster.	ECS VCI	Required
node-problem-detector	The node-problem-detector add-on is a node health monitoring add-on, which is used to detect node anomalies and report the anomalies to kube-apiserver.	ECS	Optional
log-collector	A log collection add-on for clusters. It collects cluster logs and sends them to Torch Log Service (TLS).	ECS VCI	Optional
event-collector	The event-collector add-on collects events from various cluster resources and sends them to TLS.	ECS VCI	Optional
prometheus-agent	The prometheus-agent add-on is a monitoring add-on of VMP, which collects and sends metrics of Kubernetes clusters to VMP. tip VMP officially began charging fees on May 29, 2023. If you install this add-on in the cluster, additional fees will be incurred. For details, see Billing model.	ECS VCI	Recommended
Prometheus-adapter	The prometheus-adapter is an add-on designed for VMP. It converts Prometheus metrics into the format required by the custom metric API.	ECS VCI	Optional
apmplus-opentelemetry-collector	The apmplus-opentelemetry-collector add-on is an application performance monitoring add-on, which collects and sends OpenTelemetry data and Prometheus metrics to APMPlus.	ECS	Optional
apmplus-server-agent	The apmplus-server-agent add-on collects monitoring data for services and hosts and sends the data to the application performance monitoring service.	ECS	Optional

Scheduling

Add-on name	Add-on description	Deployment method	Installation recommendations
scheduler-plugin	The scheduler-plugin functions as an expansion scheduler add-on for clusters, which expands the scheduling capabilities and policies of the native kube-scheduler. tip This is a system-managed add-on that prevents issues such as add-on faults and unsatisfactory performance caused by user-side factors. This reduces user maintenance costs and enhances add-on stability.	Users do not need to be concerned with this.	Required
vci-virtual-kubelet	The vci-virtual-kubelet add-on achieves the scheduling and use of VCI instances through virtual nodes. Virtual nodes are created by default after the add-on is deployed. tip This is a system-managed add-on, which prevents issues such as add-on faults and unsatisfactory performance caused by user-side factors. This reduces user maintenance costs and enhances add-on stability.	Users do not need to be concerned with this.	Recommended
cluster-autoscaler	The cluster-autoscaler add-on automatically scales cluster nodes. tip This is a system-managed add-on, which prevents issues such as add-on faults and unsatisfactory performance caused by user-side factors. This reduces user maintenance costs and enhances add-on stability.	Users do not need to be concerned with this.	Optional
cronhpa	The cronhpa add-on is used to achieve the scheduled scaling of workloads.	ECS VCI	Optional
katalyst	The katalyst add-on provides fine-grained resource management capabilities, such as topology reporting in RDMA scenarios and core binding in Non-Uniform Memory Access (NUMA)-aware scheduling. tip This add-on is currently in public beta. This add-on can be installed only in clusters of the following versions: Kubernetes v1.24: v1.24.10-vke.12 and later Kubernetes v1.26: v1.26.10-vke.13 and later Other later Kubernetes versions	ECS	Optional
keda	The keda add-on provides an autoscaler driven by Kubernetes events to enhance the capability of elastic scaling. tip The capabilities provided by the keda add-on are generally consistent with those provided by the open source community. For details, see the KEDA project website.	ECS VCI	Optional
karpenter	The karpenter add-on implements event-driven scale-out for cluster nodes in real time, which offers improved elastic performance.	ECS	Optional
aibrix	The aibrix functions as a solution suite for generative artificial intelligence (GenAI) inference infrastructure. It provides capabilities such as GPU workload autoscaling, with an emphasis on deep integration with the inference engine to provide advanced features and achieve ultimate optimization.	ECS VCI	Optional

DNS

Add-on name	Add-on description	Deployment method	Installation recommendations
core-dns	The core-dns add-on provides service discovery and DNS resolution services in clusters.	ECS VCI	Required
node-local-dns	The node-local-dns add-on provides the local DNS cache proxy to enhance the DNS service discovery performance and stability of the cluster.	ECS	Optional

Security

Add-on name	Add-on description	Deployment method	Installation recommendations
security-scan	An add-on that provides security baseline inspection for clusters. After the add-on is uninstalled, all inspection data is automatically deleted.	ECS	Optional
application-inspector	The application-inspector add-on is used to inspect cluster applications.	ECS VCI	Optional
pod-identity-webhook	The pod-identity-webhook add-on is an auxiliary add-on for the IAM Roles for Service Accounts (IRSA) feature, providing automated assistance for using IRSA capabilities in clusters. tip This is a system-managed add-on, which prevents issues such as add-on faults and unsatisfactory performance caused by user-side factors. This reduces user maintenance costs and enhances add-on stability.	ECS	Optional
migrate-plugin	The migrate-plugin add-on is used to back up and restore application data in VKE clusters. It can back up container application metadata and PV data to TOS and quickly restore data to a specified cluster based on backup data.	ECS	Optional
varmor	The varmor is a container security sandbox add-on designed for cloud-native environments. It performs fine-grained permission management on container runtime behavior. This restricts container access permissions at the kernel level, thereby improving the overall security defense level of clusters.	ECS	Optional

Image

Add-on name	Add-on description	Deployment method	Installation recommendations
cr-credential-controller	The cr-credential-controller add-on enables container clusters to pull private images from CR without credentials when you create applications.	ECS VCI	Optional
p2p-accelerator	[Public beta] The p2p-accelerator add-on accelerates concurrent container image pulling on a large scale by using the P2P technology.	ECS VCI	Optional
image-accelerator	The image-accelerator add-on is a container image acceleration add-on, which achieves container image lazy loading by using Nydus to accelerate image pulling during pod creation.	ECS VCI	Optional

GPU

Add-on name	Add-on description	Deployment method	Installation recommendations
nvidia-device-plugin	The nvidia-device-plugin add-on is a GPU device management add-on that enables the use of NVIDIA GPU devices in containers.	ECS	Optional
mgpu	The add-on for sharing GPU resources among containers by isolating memory and computing power between containers.	ECS	Optional

Core

Add-on name	Add-on description	Deployment method	Installation recommendations
cloud-controller-manager	The cloud-controller-manager is a core network add-on for clusters. It performs lifecycle management on cloud resources, such as nodes and load balancers. tip This is a system-managed add-on, which prevents issues such as add-on faults and unsatisfactory performance caused by user-side factors. This reduces user maintenance costs and enhances add-on stability.	Users do not need to be concerned with this.	Required