이 페이지에서는 GKE Autopilot으로 마이그레이션하여 혜택을 누릴 수 있는 사용률이 낮은 노드를 사용하여 Google Kubernetes Engine(GKE) Standard 클러스터를 식별함으로써 잠재적으로 클러스터 운영 비용을 줄이는 방법을 보여줍니다.
Autopilot에서 가능한 비용 절감
GKE Standard 클러스터에서는 실행 중인 워크로드에서 컴퓨팅 리소스를 얼마나 효율적으로 사용하는지와 상관없이 각 노드를 지원하는 Compute Engine 가상 머신(VM) 인프라에 대한 요금이 청구됩니다.
비용 효율적인 표준 클러스터를 실행하려면 빈 패킹과 같은 효율적인 리소스 소비 방법을 사용하여 리소스 사용량을 지속적으로 관리해야 합니다.
Autopilot에서는 실행 중인 워크로드가 요청하는 컴퓨팅 리소스(예: CPU 또는 메모리)에 대한 요금이 청구됩니다. 이 결제 모델을 사용하면 워크로드 리소스 사용 효율성을 모니터링하고 관리할 필요가 없습니다. 자세한 가격 정보는 Autopilot 모드 가격 책정을 참조하세요.
GKE는 실행 중인 표준 클러스터를 모니터링하여 노드의 낮은 사용률을 감지하고 마이그레이션을 통해 잠재적으로 운영 비용을 줄일 수 있는 경우 Autopilot 클러스터로 마이그레이션하는 것이 좋습니다. GKE는Google Cloud에서 리소스 사용에 대한 통계와 추천을 제공하는 서비스인 추천자를 통해 이러한 추천을 제공합니다.
사용률이 낮은 Standard 클러스터 식별
GKE는 사용률이 낮은 Standard 클러스터를 식별하는 데 도움이 되는 통계와 추천을 생성합니다. 이러한 통계를 보려면 Google Cloud 콘솔, gcloud CLI 또는 Recommender API를 사용하여 다음 파라미터를 지정합니다.
[[["이해하기 쉬움","easyToUnderstand","thumb-up"],["문제가 해결됨","solvedMyProblem","thumb-up"],["기타","otherUp","thumb-up"]],[["이해하기 어려움","hardToUnderstand","thumb-down"],["잘못된 정보 또는 샘플 코드","incorrectInformationOrSampleCode","thumb-down"],["필요한 정보/샘플이 없음","missingTheInformationSamplesINeed","thumb-down"],["번역 문제","translationIssue","thumb-down"],["기타","otherDown","thumb-down"]],["최종 업데이트: 2025-09-04(UTC)"],[],[],null,["# Identify clusters for Autopilot migration\n\n[Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview) [Standard](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n\n*** ** * ** ***\n\nThis page shows you how to potentially reduce your cluster operating costs by\nidentifying Google Kubernetes Engine (GKE) Standard clusters with\nunderutilized nodes that would benefit from a migration to GKE\nAutopilot.\n\nPotential cost savings in Autopilot\n-----------------------------------\n\nIn GKE Standard clusters, you're billed for the\nCompute Engine virtual machine (VM) infrastructure that backs each node,\nregardless of how efficiently your running workloads *use* the compute resources.\nContinually managing your resource usage using efficient resource consumption methods such as *bin-packing*\nis essential to run cost-efficient Standard clusters.\n\nIn Autopilot, you're billed for the compute resources, such as CPU or\nmemory, that your running workloads request. This billing model removes the need\nfor you to monitor and manage workload resource usage efficiency. For detailed\npricing information,\nsee [Autopilot mode pricing](/kubernetes-engine/pricing#autopilot-mode).\n\nGKE monitors your running Standard clusters to detect\nnode underutilization and recommends migrating to Autopilot clusters\nif the migration would potentially reduce your operating costs. GKE\nprovides these recommendations through [Recommender](/recommender/docs/overview),\na service that provides insights and recommendations for using resources on\nGoogle Cloud.\n\nIdentify underutilized Standard clusters\n----------------------------------------\n\nGKE generates insights and recommendations that help you to\nidentify underutilized Standard clusters. To view these insights, use\nthe Google Cloud console, the gcloud CLI, or the\nRecommender API and specify the following parameters:\n\n- **Recommender type:** `google.container.DiagnosisInsight`\n- **Subtype** : `CLUSTER_AUTOPILOT_OPTIMIZATION`\n\nFor instructions, see\n[View insights and recommendations](/kubernetes-engine/docs/how-to/optimize-with-recommenders#view-insights-recs).\n\nHow GKE identifies clusters for potential Autopilot migration\n-------------------------------------------------------------\n\nGKE uses the following signals to determine whether a\nStandard mode cluster is underutilized. The cluster must satisfy\nthe following signal for GKE to generate an insight:\n\n### Decide whether to migrate to Autopilot\n\nAutopilot clusters shift node management responsibilities to\nGKE. Before you migrate your workloads to an Autopilot\ncluster, review the following resources to decide whether Autopilot\nfits your configurability and functionality needs:\n\n- [High-level comparison of cluster modes](/kubernetes-engine/docs/concepts/choose-cluster-mode)\n- [Detailed comparison of Autopilot and Standard](/kubernetes-engine/docs/resources/autopilot-standard-feature-comparison)\n- [Autopilot security overview](/kubernetes-engine/docs/concepts/autopilot-security)\n- [About GKE Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview)\n\nIn most production environments, we recommend Autopilot clusters.\n\n### Limitations\n\nIf recent metrics from a GKE Standard cluster\naren't available, you might not see an insight for that cluster.\n\nAct on the recommendation\n-------------------------\n\nIf you see this recommendation, contact Cloud Customer Care for more information\nabout how your environment might benefit from a migration to Autopilot\nclusters. Contact your account representative or see\n[contact Cloud Customer Care](/contact).\n\nIf you decide to migrate from Standard, check workload compatibility\nwith Autopilot and prepare for a migration by following the\nguidance in\n[Prepare to migrate to Autopilot form Standard](/kubernetes-engine/docs/how-to/prepare-migrate-cluster-mode).\n\nWhat's next\n-----------\n\n- [Optimize your usage of GKE with insights and recommendations](/kubernetes-engine/docs/how-to/optimize-with-recommenders)."]]