本頁面說明如何使用 Google Cloud Managed Service for Prometheus,為 GKE 連結的叢集設定代管資料收集作業。這項功能適用於執行 Kubernetes 1.25 以上版本的 Google Cloud 叢集。如果客戶先前已在 Kubernetes 1.24 版叢集上註冊這項功能的搶先體驗計畫,該版本也支援這項功能。
這項功能可讓您使用 PodMonitoring 資源,選擇要擷取到 Cloud Monitoring 的指標。然後在 Google Cloud 控制台中查看這些指標。
限制
平台版本低於 1.23.0-gke.2 時,無法使用這項功能
這項功能需要 Linux 工作負載。
設定代管資料收集
如要使用 Managed Service for Prometheus 設定代管資料收集作業,請完成下列步驟:
[[["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:2025-07-31 (世界標準時間)。"],[],[],null,["This page explains how you can set up managed data collection with\n[Google Cloud Managed Service for Prometheus](/stackdriver/docs/managed-prometheus) for\nGKE attached clusters. This feature is available on Google Cloud clusters\nrunning Kubernetes version 1.25 or later. It is also supported on Kubernetes\nclusters at version 1.24 for customers who had earlier signed up for a\npreview of this feature on that version.\n\nWith this feature, you can use a PodMonitoring resource to choose the metrics\nthat you want to ingest into Cloud Monitoring. You can then view these\nmetrics in the Google Cloud console.\n\nLimitations\n\n- This feature is not available on platform versions earlier than 1.23.0-gke.2\n\n\u003c!-- --\u003e\n\n- This feature requires a Linux workload.\n\nSet up managed data collection\n\nTo set up managed data collection with Managed Service for Prometheus, complete the following steps:\n| **Note:** Starting with Kubernetes version 1.28, manual policy binding to authorize service accounts for data collection is no longer necessary. The required permissions are now automatically granted to this service account. You can therefore disregard step 1 in this section.\n\n1. Authorize the `gmp-system/collector` and `gmp-system/rule-evaluator` service\n accounts to write your metrics:\n\n gcloud projects add-iam-policy-binding \u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e \\\n --member=\"serviceAccount:\u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e.svc.id.goog[gmp-system/collector]\" \\\n --role=roles/monitoring.metricWriter\n gcloud projects add-iam-policy-binding \u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e \\\n --member=\"serviceAccount:\u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e.svc.id.goog[gmp-system/rule-evaluator]\" \\\n --role=roles/monitoring.viewer\n gcloud projects add-iam-policy-binding \u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e \\\n --member=\"serviceAccount:\u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e.svc.id.goog[gmp-system/rule-evaluator]\" \\\n --role=roles/monitoring.metricWriter\n\n Replace \u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e with your Google Cloud\n project ID.\n\n Adding this IAM binding fails unless at least one\n cluster has been created in your Google Cloud project. This is because the\n workload identity pool it refers to\n (\u003cvar translate=\"no\"\u003eGOOGLE_PROJECT_ID\u003c/var\u003e`.svc.id.goog`) isn't provisioned until\n cluster creation.\n2. Use the\n [`gcloud container attached clusters register` command](/sdk/gcloud/reference/container/attached/clusters/register)\n to register the cluster, and include the optional\n `--enable-managed-prometheus` flag.\n\n3. To configure which workload metrics you want to ingest, set up a\n [PodMonitoring](https://github.com/GoogleCloudPlatform/prometheus-engine/blob/v0.4.3-gke.0/doc/api.md#podmonitoring)\n resource. For an example, see\n [Configure a PodMonitoring resource](/stackdriver/docs/managed-prometheus/setup-managed#gmp-pod-monitoring)\n in the managed data collection with Managed Service for Prometheus documentation.\n\nAfter you enable managed data collection with Managed Service for Prometheus, GKE attached clusters creates the\n`gmp-system` namespace. Don't modify or deploy anything in this namespace.\nGKE attached clusters also creates the `gmp-public` namespace. Managed Service\nfor Prometheus uses the `gmp-public` namespace to deploy the\n[OperatorConfig](https://github.com/GoogleCloudPlatform/prometheus-engine/blob/v0.4.3-gke.0/doc/api.md#operatorconfig)\nresource.\n\nUpdate your Prometheus configuration\n\nTo disable or re-enable Prometheus data collection, use the\n[`gcloud container attached clusters update` command](/sdk/gcloud/reference/container/attached/clusters/update). When updating the cluster, include either the\n`--disable-managed-prometheus` or `--enable-managed-prometheus` optional flag.\n\nWhat's next\n\n- Discover how to [Configure a query user interface](/stackdriver/docs/managed-prometheus/query) in the managed data collection with Managed Service for Prometheus documentation.\n- Learn about [Managed rule evaluation and alerting](/stackdriver/docs/managed-prometheus/rules-managed) in the Google Cloud Managed Service for Prometheus documentation."]]