Standard モードでは、クラスタとノードのすべての構成設定を管理します(特性を共有するノードプールの管理も含みます)。責任共有モデルでは、引き続き Google がコントロール プレーンを管理しますが、ユーザーがノードを構成する必要があります。ユーザーが管理する設定には次のようなものがあります。
Standard モード: Pod がノードで実行されているかどうかにかかわらず、各ノード上のコンピューティング リソースに対して料金が発生します。未使用のリソースには料金がかかるため、ワークロードのスケジューリングを管理して、ノードでのリソースの無駄を最小限に抑える必要があります。詳しくは、Standard の料金をご覧ください。
Standard クラスタで一貫したレベルのリソース使用効率を確保するには、クラスタの状態を継続的にモニタリングする必要があります。Autopilot クラスタでは、GKE がモニタリングと管理を行います。
Standard クラスタでは、ノード上の未使用のコンピューティング リソースに対して料金が発生します。ビンパッキング を使用すると、可能な限り多くの Pod を各ノードに配置し、容量を無駄にしないように、これらのコストを削減できます。ビンパッキングでは、継続的なワークロード管理とスケジュールのカスタマイズが必要です。Autopilot クラスタでは、ワークロードで使用されるリソースに対してのみ料金が発生するため、ワークロードをビンパッキングする必要はありません。
Autopilot ではなく Standard を使用する場合
ほとんどのワークロードには Autopilot の使用をおすすめしますが、事前構成済みの強化されたクラスタ構成やデフォルトのクラスタ構成には、Autopilot で対応できない特有の要件がある場合があります。次のシナリオでは、Autopilot モードではなく Standard モードの使用を検討してください。
[[["わかりやすい","easyToUnderstand","thumb-up"],["問題の解決に役立った","solvedMyProblem","thumb-up"],["その他","otherUp","thumb-up"]],[["わかりにくい","hardToUnderstand","thumb-down"],["情報またはサンプルコードが不正確","incorrectInformationOrSampleCode","thumb-down"],["必要な情報 / サンプルがない","missingTheInformationSamplesINeed","thumb-down"],["翻訳に関する問題","translationIssue","thumb-down"],["その他","otherDown","thumb-down"]],["最終更新日 2025-09-01 UTC。"],[],[],null,["# Choose a GKE mode of operation\n\n[Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview) [Standard](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n\n*** ** * ** ***\n\nThis page helps you to choose the Google Kubernetes Engine (GKE) *mode of operation*\nthat's the best fit for your workloads. This page is intended for platform\nadministrators who are considering GKE as a managed Kubernetes\nprovider and want to discover the options available in Google Cloud. If\nyou want to learn about whether GKE as a *platform* is the best\nchoice for your containerized applications, refer to the\n[GKE Overview](/kubernetes-engine/docs/concepts/kubernetes-engine-overview) and [GKE and Cloud Run](/kubernetes-engine/docs/concepts/gke-and-cloud-run).\n\nGKE offers the following modes of operation for clusters:\n\n- **Autopilot mode (recommended)**: GKE manages the underlying infrastructure such as node configuration, autoscaling, auto-upgrades, baseline security configurations, and baseline networking configuration.\n- **Standard mode**: You manage the underlying infrastructure, including configuring the individual nodes.\n\nYou can't convert a cluster from Standard to Autopilot after\ncluster creation. We recommend that you read this page and, optionally, read the\n[Autopilot and Standard comparison](/kubernetes-engine/docs/resources/autopilot-standard-feature-comparison)\nso that you can make an informed choice.\n\nWhy GKE Autopilot mode\n----------------------\n\nGoogle manages most of the infrastructure in a GKE\nAutopilot cluster, providing a more *managed* Kubernetes experience\nthan GKE Standard mode. The default configuration of\nAutopilot clusters is optimized for most production workloads.\nGKE Autopilot implements many Kubernetes best practices\nfor security, scalability, and cost optimization by default.\n\n**In most cases, we recommend running your production workloads on\nAutopilot.**\n\nAutopilot provides a default configuration that has benefits such as\nthe following:\n\n- **Cost efficiency**: You only pay for the compute resources that your workloads use while running. You don't pay for unused capacity on your nodes, system Pods, operating system costs, or unscheduled workloads.\n- **Automation**: Google manages nodes, creating new nodes for your apps and configuring automatic upgrades and repairs. GKE automatically scales nodes and workloads based on traffic.\n- **Improved security posture and reliability**: Autopilot clusters enable many GKE security settings and Kubernetes best practices by default. GKE automatically applies security patches to your nodes when available.\n\nFor a full list of benefits in GKE Autopilot, refer to\n[About GKE Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview).\n\nWhy GKE Standard mode\n---------------------\n\nIn Standard mode, you manage every configuration setting in your\ncluster and nodes, including managing groups of nodes called *node pools* that\nshare characteristics. Under the\n[shared responsibility model](/kubernetes-engine/docs/concepts/shared-responsibility),\nGoogle still manages your control plane, but you must configure your nodes.\nSettings that you manage yourself include the following:\n\n- **Node pools**: You create and manage groups of nodes that have similar configuration settings.\n- **Security** : GKE Standard clusters have default hardening measures applied, but many GKE security features are not enabled by default, such as [Workload Identity Federation for GKE](/kubernetes-engine/docs/concepts/workload-identity) and [Shielded GKE Nodes](/kubernetes-engine/docs/how-to/shielded-gke-nodes). You can enable these features manually and configure the settings.\n- **Scheduling** : You must monitor and design your workloads so that GKE can schedule them efficiently on your nodes to minimize unused resources (*bin-packing*).\n- **Scaling** : You must set up and configure [node auto-provisioning](/kubernetes-engine/docs/how-to/node-auto-provisioning), configure [automatic scaling](/kubernetes-engine/docs/concepts/cluster-autoscaler) settings, and ensure that your nodes don't have too many resources or too few resources.\n- **Resource management**: You must evaluate the resource needs of each workload that you run on Standard clusters to ensure that the resource requests meet the workload requirements.\n- **Version management**: Best practices such as automatic GKE version upgrades and release channel enrollment are off by default in Standard. You can configure auto-upgrades and GKE versions when you create or update the cluster.\n\nPricing differences\n-------------------\n\nThe pricing model for Autopilot is different from Standard, as\nfollows:\n\n- **Autopilot mode** : You only pay for the compute resources that your workloads use while running. You don't pay for unused resources on nodes, OS running costs, unscheduled workloads, or system workloads. For details, refer to [Autopilot pricing](/kubernetes-engine/pricing#autopilot_mode).\n- **Standard mode** : You pay for the compute resources on each node, regardless of whether Pods run on the node. You pay for unused resources, so you should manage workload scheduling to minimize resource wastage in nodes. For details, refer to [Standard pricing](/kubernetes-engine/pricing#standard_mode).\n\nEnsuring a consistent level of resource usage efficiency in Standard\nclusters requires that you constantly monitor the state of your cluster. In\nAutopilot clusters, GKE does the monitoring and\nmanagement for you.\n\nIn Standard clusters, you pay for unused compute resources on your\nnodes. You can reduce these costs by\n[*bin-packing*](/blog/products/containers-kubernetes/gke-best-practices-to-lessen-over-provisioning),\nin which you place as many Pods as possible onto each node to avoid wasted\ncapacity. Bin-packing requires constant workload management and scheduling\ncustomization. Autopilot clusters eliminate the need for you to\nbin-pack your workloads because you only pay for the resources that your\nworkloads use.\n\nWhen to use Standard instead of Autopilot\n-----------------------------------------\n\nWhile we recommend using Autopilot for most workloads,\nyou might have specific requirements that Autopilot can't meet due to\nthe pre-configured hardening or the default cluster configuration. You should\nconsider using Standard mode over Autopilot mode in the\nfollowing scenarios:\n\n- You require granular control over your cluster and node configuration,\n including the ability to directly connect to your nodes using SSH.\n\n | **Note:** You can still get shell access to your containers in Autopilot clusters that have the `sh` binary, but not your nodes, because GKE manages Autopilot nodes. We recommend that you disable node SSH access in GKE Standard unless required.\n- You want to install or modify software running on the nodes themselves, such\n as changing the node operating system.\n\n- You want to\n [customize the node system configuration](/kubernetes-engine/docs/how-to/node-system-config),\n such as by setting\n [Linux sysctls](https://www.kernel.org/doc/Documentation/sysctl/net.txt).\n\n- You want to perform actions that Autopilot restricts, for example,\n running workloads in GKE-managed namespaces such as\n `kube-system`. We recommend that you don't deploy workloads in these\n namespaces.\n\n- You want to use specific GKE features that are only available\n on Standard, such as\n [Cloud TPU](/kubernetes-engine/docs/concepts/tpus).\n\n- You want to test alpha features in open source Kubernetes.\n\n- You want to provision additional unused capacity in your cluster.\n\nUnless you have specific requirements such as these, we recommend that you try\nAutopilot for your workloads. For an interactive walkthrough that sets\nup an Autopilot cluster and exposes a `hello-world` application, go to\nthe Autopilot walkthrough in the Google Cloud console:\n\n[Go to walkthrough](https://console.cloud.google.com/welcome?walkthrough_id=gke_autopilot)\n\nWhat's next\n-----------\n\n- [View a detailed capability comparison between Autopilot and Standard](/kubernetes-engine/docs/resources/autopilot-standard-feature-comparison).\n- [Create an Autopilot cluster](/kubernetes-engine/docs/how-to/creating-an-autopilot-cluster).\n- [Learn about GKE cluster architecture](/kubernetes-engine/docs/concepts/cluster-architecture).\n- [Learn more about Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview)."]]