Volcengine Kubernetes Engine
Volcengine Kubernetes Engine
Volcengine Kubernetes Engine
Document
Volcengine Kubernetes Engine
What's New
History feature release notes
Kubernetes version release notes
Containerd runtime release notes
Add-on release notes
Monitoring add-ons
Scheduling add-ons
DNS add-ons
Security add-ons
GPU add-ons
Sandbox add-on
Security announcements
Product Announcement
Historical announcements
Overview
Getting started
User guide
Clusters
Creating clusters
Connecting to clusters
Upgrading clusters
Node pool
Creating node pools
Scaling out/in node pools
Nodes
Authorization management
Image
Namespaces
Networking
Cluster network
Container Network
Service
LoadBalancer
CLB
Ingress
APIG Ingress
ALB Ingress
Customize ALB Ingress forwarding rules
Certificate and security
CLB Ingress
Nginx Ingress
Gateway API
Workloads
Creating workloads
Configure workloads
Pods
Configuration management
Image acceleration
Storage
FSX client
EBS Volumes
TOS Volumes
NAS volumes
FSX access methods
NFS access methods
EFS volumes
vePFS volumes
FSX access methods
vePFS Client access methods
NFS access methods
CloudFS volumes
Local volumes
Autoscale
Node autoscaler
Workload autoscaler
Horizontal Pod Autoscaler (HPA)
Cron Horizontal Pod Autoscaler (CronHPA)
Intelligent Horizontal Pod Autoscaler (IHPA)
Horizontal Pod Autoscaler based on event-driven (KEDA)
Add-on management
Network add-ons
Storage add-ons
Monitoring add-ons
Scheduling add-ons
DNS add-ons
GPU add-ons
Core add-ons
Sandbox add-on
Upgrading add-ons
Heterogeneous computing
NVIDIA GPU
mGPU scheduling
RDMA high-performance computing
Scheduling management
Topology-aware scheduling
QoS-aware scheduling
Observability
Log management
Basic observation
Setup & Configuration
Cluster observation
Storage observation
Other observation
Using metrics
Full stack observation
Alert management
Diagnosis and inspection
Node check and self-healing
Application center
Backup center
Stability
Large-scale cluster management
Best practices
Authorization
Clusters
Nodes and node pools
Networking
Storage
Scheduling
Observability
Solution
Security and compliance
API reference
Cluster management
NodePool management
Node management
Virtual nodes management
Add-on management
Tag management
Access policy management
Instance images management
Cloud native observability
Inspection and diagnosis
Task management
Volcengine Container Instance
Appendix
SDK reference
Development guide
Terraform
FAQs
General FAQs
Cluster FAQ
cluster management
cluster network
cluster upgrade
Node and Node pool FAQs
Authorization FAQs
Workload FAQs
Service and Ingress FAQs
Storage FAQs
Heterogeneous computing FAQs
Observability FAQs
Quota FAQs
Add-on FAQs
Cloud-native AI
ServingKit
LLM applications
ComfyUI application
Deploying the Al inference application through Helm
Deepseek practices
Qwen practices
Kimi practice
Seed practices
Best practices
DeepSeek model deployment
GLM model deployment
MiniMax model deployment
Qwen model deployment
Kimi model deployment
Volcengine Container Instance
What's new
Historical feature release notes
Product Announcement
Overview
Billing
User guide
Integrating with VCI
Virtual node
VKE cluster uses virtual nodes
Self-managed/third-party cluster uses virtual node
Creating an instance
General purpose instance
GPU-accelerated computing instance
VCI Agent Sandbox
Startup acceleration
E2B-based management
Upgrade guide
Networking
Monitoring
Scheduling
Container configuration
VCI Image Cache
Best practices
Security and compliance
FAQs
- Documentation
Volcengine Kubernetes EngineVolcengine Container InstanceWhat's newHistorical feature release notesHistorical release notes (2024)
Copy page
Download PDFHistorical feature release notes
Historical release notes (2024)
Copy pageDownload PDFHistorical release notes (2024)
This topic describes the updates to the features of VCI released in 2024.
Warning
Note
- Region refers to the region where the feature was first released. The release regions coming up afterward will not affect the historical release notes. For the regions where features are actually available, please refer to the information on the console pages.
- The historical records of beta testing will not change after beta features are officially released to the public. The iteration information of features is always contained in the latest release notes.
December 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Enhanced VCI scheduling capability | The scheduling capability of VCI is enhanced to specify multiple instance families at the same time. This helps you reduce concern about VCI resource limitations, such as instance family and subnet, and enables automatic resource scheduling based on the distribution of resources, such as instance family and subnet, without modifying business pods. | China (Beijing) | 2024-12-10 | |
| | China (Shanghai) | 2024-12-09 | |
| | China (Guangzhou) | 2024-12-04 | |
Added CPU metrics | The CPU metric container_cpu_load_average_10s is added. This metric allows you to obtain the average CPU load of a container in the last 10 seconds. It helps you understand the CPU resource usage of processes in the container and plan resources appropriately. | China (Beijing) | 2024-12-10 | |
| | China (Shanghai) | 2024-12-09 | |
| | China (Guangzhou) | 2024-12-04 | |
Incremental creation of image caches | In the API used to create an image cache, a new parameter has been added, which allows you to specify the image in an existing image cache as the base image to create a new image cache. This reduces redundant download operations for existing image layers and accelerates the creation and use of image caches. | China (Beijing) | 2024-12-10 | |
| | China (Shanghai) | 2024-12-09 | |
| | China (Guangzhou) | 2024-12-04 | |
November 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Configuration of Mirror image repositories | You can configure the address of the Mirror image repository. When you pull images, the requested address will be replaced with the Mirror repository address. When using a public overseas image, you can directly pull the corresponding image without changing the image address. This meets cross-region business requirements. | China (Beijing) | 2024-11-06 | |
| | China (Shanghai) | 2024-11-07 | |
| | China (Guangzhou) | 2024-11-05 | |
Scheduling pods to VCI by using annotations in self-built or cross-cloud clusters | Self-built or cross-cloud Kubernetes clusters support scheduling pods to VCI through the annotation vke.volcengine.com/burst-to-vci. This improves the efficiency of using VCI for self-built and cross-cloud clusters, and maintains consistency with the usage method of VKE clusters. | China (Beijing) | 2024-11-06 | |
| | China (Shanghai) | 2024-11-07 | |
| | China (Guangzhou) | 2024-11-05 | |
Image caches support the instant availability feature of EBS. | [Beta · Apply for Trial] Image caches can use the instant availability feature of EBS, which shortens the time from creation to availability, reduces business waiting periods, and improves resource turnover efficiency. | China (Beijing) | 2024-11-06 | |
| | China (Shanghai) | 2024-11-07 | |
| | China (Guangzhou) | 2024-11-05 | |
October 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
New GPU metrics | GPU-accelerated instances support the DCGM_FI_PROF_NVLINK_* and DCGM_CUSTOM_XID_ERRORS_* metrics. This provides GPU users with more comprehensive metrics monitoring capabilities. | China (Beijing) | 2024-10-16 | |
| | China (Shanghai) | 2024-10-15 | |
| | China (Guangzhou) | 2024-10-10 | |
September 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
In-place restart of VCI pods when partial container metadata is updated | [Beta · Apply for Trial] VCI supports in-place upgrade of pods when you update the container image, as well as the annotation or the label of the metadata for the container. In business update scenarios, pod upgrades and business recovery are accelerated. When some containers are being upgraded, others can continue running, ensuring business continuity. | China (Beijing) | 2024-09-09 | |
| | China (Shanghai) | 2024-09-05 | |
| | China (Guangzhou) | 2024-09-04 | |
Instance families supported by RCI are specific to the availability zone level | [Beta · Apply for Trial] When you select an availability zone on the RCI purchase page, instance families that are unavailable in the availability zone are dimmed. This helps you quickly purchase RCIs and improves user experience. | China (Beijing) | 2024-09-11 | RCI |
| | China (Shanghai) | 2024-09-10 | |
| | China (Guangzhou) | 2024-09-12 | |
Binding a VCI to multiple EIPs | [Beta · Apply for Trial] Multiple EIPs can be bound to a single VCI. You can replace a multi-ISP BGP EIP with multiple single-ISP EIPs, which reduces business costs. | China (Beijing) | 2024-09-03 | No |
| | China (Shanghai) | 2024-09-03 | |
| | China (Guangzhou) | 2024-08-29 | |
VCI adapting to FlexPL to configure additional performance | [Beta · Apply for Trial] VCI supports PVs of the FlexPL cloud disk type with extra performance configurations. Cloud disks of the FlexPL type with extra performance configurations can provide higher IOPS and throughput performance. | China (Beijing) | 2024-09-03 | |
| | China (Shanghai) | 2024-09-03 | |
| | China (Guangzhou) | 2024-08-29 | |
Ignoring the resource requirements of specific containers during automatic adjustment of pod instance specifications | When VCI adjusts pod instance specifications based on container limits and requests, the resource requirements of specific containers can be ignored. In this way, exclusive resources are not allocated to some containers that do not affect your business in certain scenarios. | China (Beijing) | 2024-09-11 | |
| | China (Shanghai) | 2024-09-10 | |
| | China (Guangzhou) | 2024-09-12 | |
Optimized automatic cleanup policies for image caches | [Beta · Apply for Trial] The default retention period for a manually created image cache has changed from permanent retention to 15 days. In addition, an API is provided to update image cache retention periods. This helps you reduce long-unused image caches. It ensures sufficient image cache quota for your Volcano Engine root user. It also improves service feature performance. | China (Beijing) | 2024-09-11 | |
| | China (Shanghai) | 2024-09-10 | |
| | China (Guangzhou) | 2024-09-12 | |
Specifying part of the kernel parameters for VCIs | You can specify the vm.watermark_scale_factor kernel parameter of the VCI. You can adjust the memory reclaim watermark proportionally to ensure that sufficient system resources are reserved. | China (Beijing) | 2024-09-03 | |
| | China (Shanghai) | 2024-09-03 | |
| | China (Guangzhou) | 2024-08-29 | |
August 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Integrating RCI with the Cost Center standard solution of Volcano Engine | RCI is integrated with the Cost Center standard solution of Volcano Engine, which supports features such as automatic renewal and periodic deduction. This standardizes instance deduction rules and improves the user experience. | China (Beijing) | 2024-08-13 | RCI |
| | China (Shanghai) | 2024-08-13 | |
| | China (Guangzhou) | 2024-08-12 | |
Automatically adding system labels during VCI creation | When you create a VCI, system labels, such as the cluster ID, namespace, and availability zone ID, are automatically added to the VCI. This facilitates you in querying the VCI and managing bill splitting by using these system labels. | China (Beijing) | 2024-08-13 | |
| | China (Shanghai) | 2024-08-13 | |
| | China (Guangzhou) | 2024-08-12 | |
July 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Project bill splitting | VCI integrates with the project management feature of Volcano Engine. You can specify the project a VCI belongs to and split VCI bills by project. This helps you centrally manage VCI resources and costs by project. | China (Beijing) | 2024-06-30 | |
| | China (Shanghai) | 2024-07-02 | |
| | China (Guangzhou) | 2024-06-26 | |
Supported time zone synchronization feature | You can specify time zones for virtual nodes. VCI pods can synchronize with the time zones of the virtual nodes or use hostPath to directly mount the time zone file to the container to modify the time zone of the container. This resolves the inconsistency between the time zone of the container image and the system time zone when you obtain system time for operations, such as log recording. | China (Beijing) | 2024-06-30 | |
| | China (Shanghai) | 2024-07-02 | |
| | China (Guangzhou) | 2024-06-26 | |
June 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Added general-purpose n3a instance family | [Beta · Apply for Trial] The new generation of CPU-type instance family has been released: general-purpose n3a (vci.n3a). This applies to various online business scenarios and computing-intensive scenarios. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
RCI supports renewal and more GPU instance families | [Beta · Apply for Trial] RCI supports the following features that meet general-purpose scenario requirements. These features enhance RCI capabilities.
| China (Beijing) | 2024-06-03 | RCI |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Label bill splitting | VCI supports the label bill splitting feature, which achieves bill splitting by label after you classify instances by adding labels. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Optimized automatic matching capability for VCI types | When you only specify the request for containers in VCI without specifying the limit, appropriate instance specifications can still be assigned. The automatic matching capability for VCI types is improved, which reduces the adaptation cost for migrating pods to VCI. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
In-place container image restart during update | When you update container images, you do not need to re-create pods. This achieves in-place image updates. This accelerates image updates, and business operations will not be interrupted during image updates. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
VCI supports vePFS storage | [Beta · Apply for Trial] VCI supports using vePFS-type storage services via the standard CSI method. This further expands VCI application scenarios. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
DNSCache solution | VCI supports the DNSCache solution, which provides the DNS cache proxy feature for virtual nodes. This eliminates the impact of large-scale pods on CoreDNS in clusters. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Supported network and storage performance burst for the n3i instance family | The network bandwidth, cloud disk bandwidth, and cloud disk IOPS metrics of the general-purpose n3i (vci.n3i) instance family support burst transmission. This improves the network and storage performance of VCIs and meets your requirements for high concurrency. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Configuring the startup and exit priorities of containers | You can configure the startup and exit priorities for each container in a VCI by using environment variables. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Improved GPU metrics | Eight GPU metrics are supported. This allows you to monitor and view more pod GPU status by using VMP in VCI. | China (Beijing) | 2024-06-03 | |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
Added VCI usage information in the VKE console | VCI resource usage information is added on the Overview page in the VKE console and the Overview page on the cluster details page. The pages clearly display VCI specifications, computing resource usage, and other information, which helps you collect statistics on VCI resources at the cluster level. | China (Beijing) | 2024-06-03 | No |
| | China (Shanghai) | 2024-05-30 | |
| | China (Guangzhou) | 2024-06-04 | |
May 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Ephemeral volume | VCI supports the Kubernetes-native ephemeral volumes. This enhances the consistency between VCI and native Kubernetes. | China (Beijing) | 2024-05-13 | No |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
RCI | [Beta · Apply for Trial] The RCI feature is added, which allows you to reserve resources to offset pay-as-you-go VCIs. This helps you reserve resources for a billing cycle and provides higher discounts. | China (Beijing) | 2024-05-13 | RCI |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
Image cache feature supporting CRDs | [Beta · Apply for Trial] You can create and manage VCI image caches by using the Kubernetes CRD, which automatically updates them when Volcano Engine CR images are updated. This simplifies cache use and maintenance, and avoids frequent manual updates. | China (Beijing) | 2024-05-13 | |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
Log-collector of VCI allows you to modify the default field _raw_ of TLS. | VCI updates the TLS collection agent to the latest version. You can add custom parameters to the log-collector add-on to replace the default field _raw_ in TLS. | China (Beijing) | 2024-05-13 | No |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
Pod IP addresses remaining unchanged during workload upgrades | You can enable and configure fixed IP addresses for VCI by using the vpc-cni-controlplane add-on. This allows you to perform operations on specific pod IP addresses, such as access control, service registration, service discovery, and log auditing. | China (Beijing) | 2024-05-13 | |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
Pod eviction and protection policies | Proactive eviction and protection policies are provided for VCI, which allow you to specify custom configurations based on their business requirements. This improves the availability of your application service. | China (Beijing) | 2024-05-13 | |
| | China (Shanghai) | 2024-05-15 | |
| | China (Guangzhou) | 2024-05-15 | |
March 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Creating instances by specifying multiple pod subnets | You can specify multiple subnets in the same availability zone to create an instance by using an annotation. VCI will automatically select the subnet that has the largest number of remaining IP addresses. This prevents VCI creation failures caused by insufficient capacity in some subnets in multi-subnet clusters. | China (Beijing) | 2024-03-21 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
GPU-accelerated gni3 instance family | [Beta · Apply for Trial] The new generation of GPU Accelerated instance family has been released: GPU Accelerated gni3 (vci.gni3). This instance family applies to large-scale deployment scenarios, such as generative AI inference, supports NVIDIA RTX™ features, and provides high-performance 3D graphics virtualization capabilities. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
VKE perceives the inventory of available resources for VCI during pod scheduling | [Beta · Apply for Trial] When scheduling pods, VKE can perceive the inventory of available resources for VCI. In this way, pods can be scheduled across availability zones and across ECS instances or VCIs in VKE based on the VCI inventory. This achieves higher resource elasticity and reduces the risk of perceived supply interruption during the use of VCI. | China (Beijing) | 2024-03-21 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
Enhanced vci-profile | More feature gates related to VCI global configurations are supported, such as privileged, hostNetwork, dnsPolicy, and hostIP. This helps you reduce VCI management costs by using vci-profile and supports more flexible configuration methods. | China (Beijing) | 2024-03-21 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
Rootfs mounting | [Beta · Apply for Trial] You can mount the rootfs of a VCI to EBS for persistent storage. In this way, after a VCI exits, related data, such as packages installed through apt-get, can still be retained persistently. | China (Beijing) | 2024-03-21 | No |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
Release of the general-purpose computing power u1 instance family in all regions | The general-purpose computing u1 instance family is available in all regions. If you do not specify an instance specification when creating a VCI, u1 is used by default in all regions. Users who are not sensitive to CPU generation can use the u1 instance type family to achieve higher cost-effectiveness and sufficient VCI inventory. | China (Beijing) | 2024-03-21 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
Creating DaemonSet-type resources | [Beta · Apply for Trial] Resources of the DaemonSet type can be created and converted into sidecar containers for injection into business VCI pods. In this way, DaemonSets can be deployed on VCIs in specific scenarios such as log collection, monitoring, and gateway, which provides higher Kubernetes-native consistency. | China (Beijing) | 2024-03-21 | No |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
VCI image caches supporting fuzzy search by cache name | [Beta · Apply for Trial] The API for querying VCI image caches has been optimized and now includes features such as fuzzy search by name and sorting by creation time and last usage time. User experience optimizations for image caches are also provided | China (Beijing) | 2024-03-21 | |
| | China (Shanghai) | 2024-03-25 | |
| | China (Guangzhou) | 2024-03-22 | |
January 2024
Feature name | Description | Region | Release date | Related documentation |
Feature name | Description | Region | Release date | Related documentation |
Image cache update | [Beta · Apply for Trial] The warm-up period of an image cache can be extended through the API used to update the image cache. This improves the user experience of using image cache capabilities. | China (Beijing) | 2024-01-29 | No |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Improved VCI metrics | Eleven GPU metrics are supported. This allows you to monitor more pod GPU status by using VMP in VCI. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
General-purpose n3i instance family | [Beta] The new generation of general-purpose instance family is released: general-purpose n3i (vci.n3i). This instance family provides you with VCI types with higher performance and lower power consumption in terms of computing, network, and storage. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Available GPU driver versions | The GPU driver version is available when you create a VCI. This enables you to specify the desired NVIDIA driver version from multiple available versions when creating GPU-accelerated instances. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Obtaining process coredump files | VCI allows you to obtain process coredump files, which facilitates viewing and analyzing coredump files to locate issues when containers unexpectedly terminate. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Vci-profile supporting non-intrusive deployment of services in VCI pods | VCI resources at the cluster or namespace level can use unified configurations and global fixed configurations. This capability reduces the need to modify workload YAML files, enables more convenient, efficient, and non-intrusive use of VCI, and avoids confusion between operation management and business management. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
New instances types of the general-purpose computing power u1 instance family | The following instance specifications are added to the general-purpose computing power u1 (vci.u1) instance family: The instance specifications of the general-purpose u1 instance family are refined and increased. This meets your business requirements for the computing resources of VCI pod instances in a more fine-grained manner.
| China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Cache FileNAS and general-purpose FileNAS | VCI supports cache-type FileNAS storage and general-purpose FileNAS storage. This applies to scenarios, such as workloads using high-speed file system interfaces for data access and storage of frequently accessed data. | China (Beijing) | 2024-01-29 | |
| | China (Shanghai) | 2024-01-29 | |
| | China (Guangzhou) | 2024-01-23 | |
Last updated: 2026.07.24 16:18:06
All rights reserved © Beijing Volcano Engine Technology Co., Ltd2024
