Update | Description | Reference |
New model for TKE native nodes | TKE native nodes support the M3 models. | |
Introduction of the industry's first memory compression capability for container scenarios | "Cold" memory is recycled and compressed based on the "WuJing" kernel. This realizes control of the pod-level compression scope, helping customers solve cluster memory bottlenecks. | - |
Support for users to use Nginx-ingress-controller to create ingress after TKE registered nodes are enabled | After registered nodes are enabled in a TKE cluster, ingress of the Nginx-Ingress-controller type is no longer blocked. Users can follow the documentation to create Nginx-ingress-controller instances through standard scheduling policies, and then use them according to the standard process. | |
Optimization of the node lifecycle management experience | Users can adjust the expected number of instances when deleting a node. |
Update | Description | Reference |
Support for importing Tencent Cloud SSM credentials using ExternalSecretOperator | ExternalSecretOperator can import key credentials, which are uniformly stored and managed in Tencent Cloud Secrets Manager (SSM), into a cluster in the form of Kubernetes native secret objects and achieve automatic synchronization of key data. This enables the unified storage and management of key lifecycles by SSM. | - |
Support for TKE registered nodes to access to public network nodes | Registered nodes are fully launched with a new public network version, allowing users to easily manage IDC nodes in a public cloud environment (without dedicated lines), quickly build Kubernetes clusters, and reduce costs. |
Update | Description | Reference |
Support for CFS Turbo in cluster storage | CFS Turbo is supported for cluster storage, suitable for bushiness with high throughput and hybrid workloads. | |
Upgrade of the native node dedicated scheduler | The native node dedicated scheduler is upgraded to enhance stability with high availability deployment, support configuring independent drain stop watermarks, and support scheduling event logs. | |
New models for TKE native nodes | TKE native nodes support the M6 and GT4 models. |
Update | Description | Reference |
GooseFS-Lite supported as a COS mounting method | GooseFS-Lite is supported as a COS mounting method, enabling higher large file read and write speeds that are not limited by local disk performance. | |
Verification for deprecated API versions in pre-upgrade cluster check | Pre-upgrade cluster check includes verification for deprecated API versions to avoid compatibility issues during upgrade, ensuring cluster stability and reliability. | |
Support for updating Management parameters of legacy nodes | Native node pools support in-place updates of Kubelet, Nameservers, Hosts, and KernelArgs (kernel) parameters through Management. | |
Support for wildcard domain name configuration for Ingress | Forwarding rules and certificates of the wildcard domain name type can be configured for CLB-type Ingress. | |
Local binding and weighting enabled by default in the externalTrafficPolicy local mode of TKE services | In the externalTrafficPolicy local mode of TKE services, CLB by default binds to the node where the pod is located and weights according to the number of pods on the node. | |
Graceful Shutdown and Graceful Deletion enabled by default in the CLB Direct Connect mode | Graceful Shutdown and Graceful Deletion are enabled by default in the CLB Direct Connect mode, enhancing network stability. |
Update | Description | Reference |
Optimization of TKE native node scheduler capabilities | The native node dedicated scheduler supports configuring independent drain stop watermarks. | |
New log collection capabilities | Based on log collection rules, users can configure multiple paths as container file paths and node file paths. | - |
Premium SSD CBS, supporting setting additional performance | Users can create a Premium SSD CBS StorageClass, which supports setting additional performance for higher performance. | |
Optimization of Nginx Ingress capabilities | Nginx Ingress supports configuring forwarding rules of the HTTP and HTTPS types through the console. |
Update | Description | Reference |
Optimization of general node experience | Error messages in the initialization process of general nodes are more precise for error locating, supporting the display of specific reasons for initialization errors. | |
| Cluster upgrade prompts are more readable, displaying the version limit details of dependent components required for the upgrade. | |
New models for TKE general nodes | Premium CVM models of the A800/H800 type are supported. | |
New capabilities for TKE native nodes | Enhanced stability: The native node dedicated scheduler is fixed for higher stability. | |
| New capability: Modification of security groups in native node pools takes effect on legacy nodes. | |
| New capability: Native node pools support modifying subnets. | |
| New capability: Native nodes support setting Tencent Cloud tags. | |
| Experience optimization: Monitoring button display is added for native node pools. | |
Update | Description | Reference |
KMS encryption for Secrets in managed/independent clusters | TKE clusters can use KMS to encrypt and decrypt various sensitive data in Secrets, such as database usernames, passwords, certificates, OAuth Tokens, and SSH KEYS, enhancing data security. | |
Cluster upgrade capability iteration | TKE Kubernetes version supports upgrading from 1.22 to 1.24. | |
Update on node/container login method for TKE | New capability: Node/Container login for TKE supports OrcaTerm. | - |
New capabilities for TKE native nodes | New capability: Native nodes support console-initiated restart. | |
| New capability: Trace scenario detection is added for RuntimeUnhealthy, supporting runtime restart for kubelet status exceptions caused by this scenario. | |
| Experience optimization: Pod running details are displayed during node draining. | |
| Experience optimization: For prepaid node pools, users can search for target nodes and check node cordon status when adjusting the node quantity. | |
| Experience optimization: Prepaid nodes support bulk renewal on the billing platform. | |
Update | Description | Reference |
Connection to a cluster through Cloud Shell | TKE has integrated Tencent Cloud Cloud Shell, allowing users to connect to a cluster with one click from the Tencent Cloud console and flexibly manage the cluster via kubectl. | |
Ingress certificate modification and upgrade | All existing certificates for the TKE service can be modified with one click on the certificate platform. | - |
New capabilities for TKE Insight | Users can view the cost trend and forecasting of TKE clusters, including the cost distribution of workloads and namespaces. | - |
New capabilities for TKE native nodes | New models: The S6, GN10X, and GN10Xp models are supported for the China site. | |
| New drivers: GPU models support the installation of 450/470/515 drivers. | |
| New capability: Enhanced security capability can. | |
| New capability: Native node pools support modifying security groups and data disks. | |
Strong verification for service activation provided for using CFS | Before the CFS component is installed in TKE, strong verification and guidance for service activation is provided to avoid scenarios where services have not been activated but resources are in use through weak API validation, resulting in billing. This enhances the linkage and billing accuracy between TKE and CFS. | |
Support for search within the namespace of log collection rules | When configuring a log source for log collection rules, users can select a namespace in a drop-down menu and search one or more namespaces. In this way, users can quickly find and select namespaces from which logs need to be collected. | - |
Update | Description | Reference |
Support for cross-region shipment of cluster audit logs and event logs | To facilitate multi-region centralized management, cluster audit logs and event logs can be shipped to regions outside of the cluster's location for centralized and unified management. | |
Upgrade of access layer components | The mode of Service component lifecycle management is upgraded to reduce anomalies during Service modification. For certain misconfigurations of Service/Ingress, TKE Service/Ingress components will opt to skip rather than stop running, preventing the risk of traffic interruption due to Service/Ingress anomalies. | - |
Launch of TKE Backup Center | TKE Backup Center provides an integrated solution for backup, recovery, and migration of containerized applications. |
Update | Description | Reference |
Nginx Ingress independent of EIP | The Nginx Ingress controller no longer depends on EIP, reducing user costs. | |
Support for formatted mounting of data disks and addition of alternative models with the same specifications during the creation of native node pools | Data disk: Native node pools support formatted mounting of Premium SSD CBS, with a recommended data disk size of at least 50 GB. Alternative model: Native node pools support configuring multiple alternative models with the same specifications, which can be set on the node pool details page. |
Update | Description | Reference |
Launch of the runtime upgrade capability | The cluster node upgrade entry also supports the minor version updates of docker and containerd components. | - |
Support for metadata customization by using cluster log collection rules | When configuring rules for collecting logs and shipping them to Tencent Cloud Log Service (CLS) in the TKE console, you can customize metadata to reduce unnecessary storage resources. |
Was this page helpful?