功能閘道是與 Kubernetes API 不同的功能啟用機制，GKE 叢集會以不同方式管理功能閘道。Alpha 版叢集會啟用所有 Kubernetes Alpha 版 API。如果您為 Alpha 叢集啟用或停用功能閘道，Kubernetes API 的啟用狀態不會受到影響。不過，部分 Kubernetes API 和 Kubernetes 功能閘道會一起運作，因此停用功能閘道可能會導致相關 API 發生問題。驗證特定功能閘道和 API 之間的連線。
[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["難以理解","hardToUnderstand","thumb-down"],["資訊或程式碼範例有誤","incorrectInformationOrSampleCode","thumb-down"],["缺少我需要的資訊/範例","missingTheInformationSamplesINeed","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間：2025-09-01 (世界標準時間)。"],[],[],null,["# Alpha clusters\n\n[Standard](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n\n*** ** * ** ***\n\nThis page provides an overview of how alpha clusters work in\nGoogle Kubernetes Engine (GKE). Access to alpha features and APIs works differently\nwith alpha clusters than with other types of GKE clusters. You\ncan [create an alpha\ncluster](/kubernetes-engine/docs/how-to/creating-an-alpha-cluster) or learn more\n[about cluster configuration\nchoices](/kubernetes-engine/docs/concepts/types-of-clusters).\n| **Warning:** Don't use alpha clusters or alpha features for production workloads. Alpha clusters expire after 30 days and don't receive security updates. You must migrate your data from alpha clusters before they expire. GKE does not automatically save data stored on alpha clusters. For more information, see [Limitations](#limitations).\n\nYou can experiment with Kubernetes alpha features by creating an *alpha\ncluster*. Alpha clusters are short-lived clusters that run stable Kubernetes\nreleases. All Kubernetes APIs, including alpha APIs, are enabled. By default,\nalpha clusters have all alpha feature gates enabled that are available with the\nKubernetes version. Some beta feature gates are enabled, depending on the\ndefault for Kubernetes. You can also selectively enable or disable specific\nalpha or beta feature gates during cluster creation. Alpha clusters are designed\nfor advanced users and early adopters to experiment with workloads that take\nadvantage of new features before those features are production-ready.\n\n### Limitations\n\nAlpha clusters have the following limitations:\n\n- Not covered by the [GKE SLA](/kubernetes-engine/sla)\n- Can't be upgraded\n - You must disable node auto-upgrade and auto-repair on alpha clusters\n- Can't be enrolled in a [release channel](/kubernetes-engine/docs/concepts/release-channels)\n- Are automatically deleted after 30 days\n- Don't receive security updates\n- Not supported with Windows Server node pools\n- Have additional [limitations for enabling feature gates with alpha clusters](#limitations-alpha-feature-gate)\n\n### Difference between alpha clusters and alpha GKE versions\n\nAlpha clusters don't necessarily run \"alpha\" versions of GKE. The\nterm *alpha cluster* means that alpha APIs are enabled regardless of the version\nof Kubernetes the cluster runs. Periodically, Google offers customers the\nability to test and validate GKE versions that are not generally\navailable. These early-access versions can be run as alpha clusters or as\nclusters without the Kubernetes alpha APIs enabled.\n\nFeature gates and alpha clusters\n--------------------------------\n\nFeature gates are a set of key-value pairs that toggle Kubernetes features. With\nalpha clusters, GKE enables the following:\n\n- **Alpha feature gates**: all feature gates at the alpha stages that are available with the Kubernetes version of your cluster\n- **Beta feature gates**: the feature gates at the beta stages that are available with the Kubernetes version of your cluster, and which are enabled by default by open source Kubernetes\n\nFor more information about which feature gates are available for a specific\nKubernetes version, and which beta features are enabled by default (`true`), see\n[Feature gates for Alpha or Beta features](https://kubernetes.io/docs/reference/command-line-tools-reference/feature-gates/#feature-gates-for-alpha-or-beta-features)\nin the Kubernetes documentation. However, you can selectively enable or disable\nspecific feature gates, modifying the default values set by GKE.\n\nFeature gates are a distinct mechanism for feature enablement from Kubernetes\nAPIs, and are managed differently by GKE clusters. With alpha\nclusters, all Kubernetes alpha APIs are enabled. If you enable or disable a\nfeature gate for an alpha cluster, the enablement of Kubernetes APIs isn't\naffected. However, some Kubernetes APIs and Kubernetes feature gates work\ntogether, so disabling feature gates might cause issues with related APIs.\nVerify the connection between specific feature gates and APIs.\n\nFor more information about how GKE works with feature gates, see\n[Feature gates](/kubernetes-engine/docs/concepts/feature-gates).\n\nTo enable a cluster with a specific feature gate configuration, use the\n`--alpha-cluster-feature-gates` flag during cluster creation with the\nGoogle Cloud CLI. When you include this flag, you can modify which alpha and beta\nfeature gates are enabled or disabled. For more information, see [Create an\nalpha\ncluster](/kubernetes-engine/docs/how-to/creating-an-alpha-cluster#create_an_alpha_cluster).\n\n### Limitations for enabling feature gates with alpha clusters\n\nWhen you enable feature gates with alpha clusters, understand the following\nadditional limitations:\n\n- You can only create alpha clusters with specific feature gates by using the gcloud CLI. Cluster creation with the Google Cloud console or Terraform isn't supported.\n- Kubernetes alpha features that require additional configuration beyond enabling the feature gate aren't supported.\n- You can selectively enable or disable alpha or beta feature gates, but all Kubernetes alpha APIs are enabled.\n\nLatest Kubernetes alpha features\n--------------------------------\n\nMost Kubernetes releases contain new alpha features that you can test in alpha\nclusters. For a full list of Kubernetes releases and the features they include,\nsee the [Kubernetes changelog](https://github.com/kubernetes/kubernetes/blob/master/CHANGELOG/README.md).\n\n### About feature stages\n\nNew Kubernetes features are introduced in four stages: early development, alpha,\nbeta, and stable.\n\nTo provide stability and production quality, GKE clusters other\nthan alpha clusters enable only features that are *beta or higher*.\nGKE doesn't enable alpha features on other types of\nGKE clusters because the features aren't production-ready or\nupgradeable.\n\nBecause GKE automatically upgrades the Kubernetes control plane\nand, by default, the worker nodes, enabling alpha features in production can\njeopardize the reliability of the cluster if there are breaking changes in a new\nversion.\n\nTo learn more about the stages of Kubernetes features, see [Alpha, Beta, and\nStable\nVersions](https://github.com/kubernetes/community/blob/master/contributors/devel/sig-architecture/api_changes.md#alpha-beta-and-stable-versions)\nin the Kubernetes documentation.\n\nTo learn more about determining which features are enabled for a given\nGKE control plane version, see [Feature\ngates](/kubernetes-engine/docs/concepts/feature-gates).\n\nWhat's next\n-----------\n\n- Learn more [about cluster configuration choices](/kubernetes-engine/docs/concepts/types-of-clusters).\n- [Create an alpha cluster](/kubernetes-engine/docs/how-to/creating-an-alpha-cluster).\n- [Read the GKE overview](/kubernetes-engine/docs/concepts/kubernetes-engine-overview)."]]
