Overview
Last updated: 2025-03-24 18:25:04
Definition
Native nodes are a new type of nodes that TKE provides for Kubernetes environments. Developed based on Tencent Cloud's technological experience in operations of ten million-core containers, native nodes provide users with native, highly stable, and fast-responding Kubernetes node management capabilities.
Benefits
Integrated with the FinOps principle, native nodes boost cost optimization of cloud resources
Provide the HouseKeeper visualized resource dashboard to improve resource utilization, reduce costs, and enhance efficiency.
Support request recommendation to avoid resource idling.
Provide dedicated dynamic scheduling capabilities with the following features:
Load balancing: Better balance the resource load on nodes based on the current load and load history.
Packing improvement: Increase the amount of schedulable node CPU and memory resources and improve the node packing rate to over 100%.
Business adjustment: Allow you to set an expected resource utilization rate and ensure continuous node scheduling to realize more centralized business resource deployment.
Native nodes provide multi-dimensional management capabilities to reduce operations workload
New Kubernetes operations model: Provide the declarative infrastructure API for users to manage nodes in the same way as workload management. You can manage nodes using the Kubernetes API, the Tencent Cloud API, or the TKE console.
Tencent’s intelligent operations system: Support operating system-, runtime-, and Kubernetes-level fault detection and automatic upgrade to reduce operations workload for users.
Based on Tencent Cloud’s cloud-native technology practices, parameters are tuned and adapted comprehensively at the operating system, runtime, and Kubernetes levels, significantly enhancing the node initialization stability.
Native Nodes vs. Normal Nodes
In general, native nodes provide all the capabilities that normal nodes have, with better performance. The following table compares the two types of nodes in details.
Module | Native Nodes | Normal Modes |
Management mode | Node HouseKeeper mode: This mode provides resource management and stable operations capabilities to assist users in decision-making. | Serverful mode: Analysis, decision-making, and action are all conducted by users. |
Declarative management of infrastructure | Not supported | |
In-place Pod configuration adjustment | Not supported | |
Custom configuration entries for kernel parameter tuning and other purposes | Not supported | |
Node self-heal | Tencent’s operating system-, Kubernetes-, and runtime-level fault detection and self-heal capabilities | NPD add-on (discontinued) |
Scheduler | Schedulers dedicated for native nodes, supporting virtual scale-outs of schedulable resources | Dynamic Scheduler and DeScheduler |
Request Recommendation | Recommendation and quick update | Not supported |
Job occupation | Supported | Not supported |
Manage Node | Kernel parameter configuration, nameserver parameter configuration, host parameter configuration, pre-request scripts, and post-request scripts | Not supported |
Billing Modes
Native nodes support multiple types of Cloud Virtual Machine (CVM) instances. You can select appropriate instances to deploy based on your application scale and business characteristics. TKE charges resources (including CPU, memory, GPU, and system disk resources) consumed by native nodes according to the node type and resource specifications.
Native nodes supported the pay-as-you-go billing mode.
Billing Mode | Pay-as-you-go |
Payment method | Postpaid (Fees are frozen upon purchase and settled hourly) |
Billing unit | USD per second |
Price | High |
Minimum use duration | Billing per second and settlement per hour; resource purchase and release allowed at any time |
Use case | Suitable for periodic computing scenarios such as transcoding, big data, and e-commerce flash sale campaigns, or tidal online service scenarios where Horizontal Pod Autoscaler (HPA) is enabled. |
Regions and Availability Zones
You can use native nodes in the following regions.
China
Region | Region Abbreviation | ||
China | Regions on the public cloud | Beijing | ap-beijing |
| | Nanjing | ap-nanjing |
| | Shanghai | ap-shanghai |
| | Guangzhou | ap-guangzhou |
| | Chengdu | ap-chengdu |
| | Chongqing | ap-chongqing |
| | Hong Kong (China) | ap-hongkong |
| | Taipei (China) | ap-taipei |
Other countries and regions
Region | Region Abbreviation | ||
Asia Pacific | Regions on the public cloud | Singapore | ap-singapore |
| | Jakarta | ap-jakarta |
| | Seoul | ap-seoul |
| | Bangkok | ap-bangkok |
| | Tokyo | ap-tokyo |
North America | | Silicon Valley | na-siliconvalley |
| | Virginia | na-ashburn |
South America | | São Paulo | sa-saopaulo |
Europe | | Frankfurt | eu-frankfurt |
References
Note
For the convenience of management, we recommend that you use a node pool to create and manage a group of native nodes with the same parameter settings.
Creating Native Nodes: You can create a native node by using the TKE console, calling the Kubernetes API, or calling the TencentCloud API.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No