이 문서에서는 Google Cloud Managed Service for Prometheus를 사용하여 Inference Optimized Gateway에서 측정항목을 수집할 수 있도록 Google Kubernetes Engine 배포를 구성하는 방법을 설명합니다. 이 문서에서는 다음을 수행하는 방법을 보여줍니다.
측정항목을 보고하도록 Inference Optimized Gateway를 설정합니다.
내보낸 측정항목을 수집하도록 Managed Service for Prometheus의 PodMonitoring 리소스를 구성합니다.
Managed Service for Prometheus 및 관리형 컬렉션을 사용하여 Inference Optimized Gateway 내보내기 도구에서 측정항목을 수집하려면 배포가 다음 요구사항을 충족해야 합니다.
클러스터가 Google Kubernetes Engine 버전 1.21.4-gke.300 이상을 실행 중이어야 합니다.
관리형 컬렉션이 사용 설정된 상태에서 Managed Service for Prometheus를 실행 중이어야 합니다. 자세한 내용은 관리형 컬렉션 시작하기를 참조하세요.
Inference Optimized Gateway는 Prometheus 형식 측정항목을 자동으로 노출하므로 별도로 설치할 필요가 없습니다. Inference Optimized Gateway 내보내기 도구가 예상 엔드포인트에서 측정항목을 내보내는지 확인하려면 다음을 수행합니다.
측정항목 스크래핑 문서에 설명된 대로 다른 터미널 세션의 curl 유틸리티를 사용하여 엔드포인트 localhost:9090/metrics에 액세스합니다.
PodMonitoring 리소스 정의
대상 검색을 위해 Managed Service for Prometheus 연산자에는 동일한 네임스페이스의 Inference Optimized Gateway 내보내기 도구에 해당하는 PodMonitoring 리소스가 필요합니다.
다음 PodMonitoring 구성을 사용할 수 있습니다.
# Copyright 2025 Google LLC## Licensed under the Apache License, Version 2.0 (the "License");# you may not use this file except in compliance with the License.# You may obtain a copy of the License at## https://www.apache.org/licenses/LICENSE-2.0## Unless required by applicable law or agreed to in writing, software# distributed under the License is distributed on an "AS IS" BASIS,# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.# See the License for the specific language governing permissions and# limitations under the License.apiVersion:monitoring.googleapis.com/v1kind:PodMonitoringmetadata:name:inference-optimized-gateway-monitoringlabels:app.kubernetes.io/name:inference-optimized-gatewayapp.kubernetes.io/part-of:google-cloud-managed-prometheusspec:endpoints:-port:metricsscheme:httpinterval:5spath:/metricsselector:matchLabels:app:inference-gateway-ext-proc
port 및 matchLabels 필드의 값이 모니터링하려는 Inference Optimized Gateway 포드의 값과 일치하는지 확인합니다.
[[["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: 2025-07-15(UTC)"],[],[],null,["# Inference Optimized Gateway\n\n\u003cbr /\u003e\n\nThis document describes how to configure your Google Kubernetes Engine deployment\nso that you can use Google Cloud Managed Service for Prometheus to collect metrics from\n\nInference Optimized Gateway. This document shows you how to do the following:\n\n- Set up Inference Optimized Gateway to report metrics.\n- Configure a PodMonitoring resource for Managed Service for Prometheus to collect the exported metrics.\n- Access a dashboard in Cloud Monitoring to view the metrics.\n\n\u003cbr /\u003e\n\nThese instructions apply only if you are using [managed collection](/stackdriver/docs/managed-prometheus/setup-managed)\nwith Managed Service for Prometheus.\nIf you are using self-deployed collection, then see the\n\n[Inference Optimized Gateway documentation](https://github.com/kubernetes-sigs/gateway-api-inference-extension/blob/main/site-src/guides/metrics.md)\n\nfor installation information.\n\nThese instructions are provided as an example and are expected to work in\nmost Kubernetes environments.\n\nIf you are having trouble installing an\napplication or exporter due to restrictive security or organizational policies,\nthen we recommend you consult open-source documentation for support.\n\nFor information about Inference Optimized Gateway, see [Inference Optimized Gateway](https://github.com/kubernetes-sigs/gateway-api-inference-extension/).\n\nPrerequisites\n-------------\n\nTo collect metrics from\n\nthe Inference Optimized Gateway exporter\n\nby using\nManaged Service for Prometheus and managed collection, your deployment must\nmeet the following requirements:\n\n- Your cluster must be running Google Kubernetes Engine version 1.21.4-gke.300 or later.\n- You must be running Managed Service for Prometheus with managed collection enabled. For more information, see [Get started with managed collection](/stackdriver/docs/managed-prometheus/setup-managed).\n\n \u003cbr /\u003e\n\n\u003cbr /\u003e\n\nInference Optimized Gateway exposes Prometheus-format metrics automatically; you do not have to install it separately. To verify that the Inference Optimized Gateway exporter is emitting metrics on the expected endpoints, do the following:\n\n\u003cbr /\u003e\n\n1. Set up port forwarding by using the following command:\n\n ```\n kubectl -n NAMESPACE_NAME port-forward POD_NAME 9090\n ```\n2. Access the endpoint `localhost:9090/metrics` by using the `curl` utility in another terminal session as described in the [Scrape Metrics document](https://github.com/kubernetes-sigs/gateway-api-inference-extension/blob/main/site-src/guides/metrics.md#scrape-metrics).\n\nDefine a PodMonitoring resource\n-------------------------------\n\nFor target discovery, the Managed Service for Prometheus Operator\nrequires a PodMonitoring resource that corresponds to the Inference Optimized Gateway exporter\nin the same namespace.\n\nYou can use the following PodMonitoring configuration: \n\n # Copyright 2025 Google LLC\n #\n # Licensed under the Apache License, Version 2.0 (the \"License\");\n # you may not use this file except in compliance with the License.\n # You may obtain a copy of the License at\n #\n # https://www.apache.org/licenses/LICENSE-2.0\n #\n # Unless required by applicable law or agreed to in writing, software\n # distributed under the License is distributed on an \"AS IS\" BASIS,\n # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n # See the License for the specific language governing permissions and\n # limitations under the License.\n\n apiVersion: monitoring.googleapis.com/v1\n kind: PodMonitoring\n metadata:\n name: inference-optimized-gateway-monitoring\n labels:\n app.kubernetes.io/name: inference-optimized-gateway\n app.kubernetes.io/part-of: google-cloud-managed-prometheus\n spec:\n endpoints:\n - port: metrics\n scheme: http\n interval: 5s\n path: /metrics\n selector:\n matchLabels:\n app: inference-gateway-ext-proc\n\nEnsure that the values of the `port` and `matchLabels` fields match those of the Inference Optimized Gateway pods you want to monitor.\n\nTo apply configuration changes from a local file, run the following command:\n\n```\nkubectl apply -n NAMESPACE_NAME -f FILE_NAME\n```\n\n\u003cbr /\u003e\n\nYou can also\n[use Terraform](/stackdriver/docs/managed-prometheus/setup-managed#terraform-scrape)\nto manage your configurations.\n\nVerify the configuration\n------------------------\n\nYou can use Metrics Explorer to verify that you correctly configured\nthe Inference Optimized Gateway exporter. It might take one or two minutes for\nCloud Monitoring to ingest your metrics.\n\nTo verify the metrics are ingested, do the following:\n\n1. In the Google Cloud console, go to the\n *leaderboard* **Metrics explorer** page:\n\n [Go to **Metrics explorer**](https://console.cloud.google.com/monitoring/metrics-explorer)\n\n \u003cbr /\u003e\n\n If you use the search bar to find this page, then select the result whose subheading is\n **Monitoring**.\n2. In the toolbar of the query-builder pane, select the button whose name is either *code* **MQL** or *code* **PromQL**.\n3. Verify that **PromQL** is selected in the **Language** toggle. The language toggle is in the same toolbar that lets you format your query.\n4. Enter and run the following query: \n\n ```\n inference_model_request_total{cluster=\"CLUSTER_NAME\", namespace=\"NAMESPACE_NAME\"}\n ```\n\n\u003cbr /\u003e\n\nView dashboards\n---------------\n\nThe Cloud Monitoring integration includes\n\nthe **Inference Optimized Gateway Prometheus Overview** dashboard.\n\nDashboards are automatically installed when you configure the integration.\nYou can also view static previews of dashboards without installing the\nintegration.\n\n\nTo view an installed dashboard, do the following:\n\n1. In the Google Cloud console, go to the **Dashboards** page:\n\n [Go to **Dashboards**](https://console.cloud.google.com/monitoring/dashboards)\n\n \u003cbr /\u003e\n\n If you use the search bar to find this page, then select the result whose subheading is\n **Monitoring**.\n2. Select the **Dashboard List** tab.\n3. Choose the **Integrations** category.\n4. Click the name of the dashboard, for example, **Inference Optimized Gateway Prometheus Overview**.\n\n\u003cbr /\u003e\n\nTo view a static preview of the dashboard, do the following:\n\n1. In the Google Cloud console, go to the\n **Integrations**\n page:\n\n [Go to **Integrations**](https://console.cloud.google.com/monitoring/integrations)\n\n \u003cbr /\u003e\n\n If you use the search bar to find this page, then select the result whose subheading is\n **Monitoring**.\n2. Click the **Kubernetes Engine** deployment-platform filter.\n3. Locate the Inference Optimized Gateway integration and click **View Details**.\n4. Select the **Dashboards** tab.\n\n\u003cbr /\u003e\n\nTroubleshooting\n---------------\n\nFor information about troubleshooting metric ingestion problems, see\n[Problems with collection from exporters](/stackdriver/docs/managed-prometheus/troubleshooting#exporter-problems) in [Troubleshooting ingestion-side problems](/stackdriver/docs/managed-prometheus/troubleshooting#ingest-problems)."]]
대시보드 페이지로 이동합니다.
통합 페이지로 이동합니다.