本文档介绍了如何在裸机上使用 Google Distributed Cloud（纯软件）的 Cloud Audit Logs。Google Distributed Cloud 使用 Kubernetes 审核日志记录功能，该功能会按时间顺序记录对集群 Kubernetes API 服务器进行的调用。审核日志适合用于调查可疑的 API 请求以及收集统计信息。如需了解 GKE On-Prem API 的审核日志记录，请参阅 Cloud API 审核日志记录。
gcloudloggingread\'logName="projects/PROJECT_ID/logs/externalaudit.googleapis.com%2Factivity" \ AND resource.type="k8s_cluster" \ AND protoPayload.serviceName="anthosgke.googleapis.com" '\--limit2\--freshness300d
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC)：2025-08-08。"],[],[],null,["This document describes how to use Cloud Audit Logs for Google Distributed Cloud\n(software only) on bare metal. Google Distributed Cloud uses [Kubernetes Audit\nLogging](https://kubernetes.io/docs/tasks/debug-application-cluster/audit/),\nwhich keeps a chronological record of calls made to a cluster Kubernetes API\nserver. Audit logs are useful for investigating suspicious API requests and for\ncollecting statistics. For information about audit logging for the\nGKE On-Prem API, see [Cloud API audit\nlogging](/kubernetes-engine/distributed-cloud/bare-metal/docs/how-to/audit-logging-api).\n| **Note:** Starting with Google Distributed Cloud release 1.9.0, Cloud Audit Logs is enabled by default. Cloud Audit Logs is automatically enabled for 1.8.x clusters that are upgraded to 1.32.400-gke.68 unless it was explicitly disabled for the 1.8.x cluster by setting `disableCloudAuditLogging` to `true`.\n\nAbout Cloud Audit Logs\n\nAudit logs are written to [Cloud Audit Logs](/logging/docs/audit) in your\nGoogle Cloud project. Writing to Cloud Audit Logs has several benefits over writing to\ndisk or capturing logs in an on-premises logging system:\n\n- Audit logs for all GKE clusters can be centralized.\n- Log entries written to Cloud Audit Logs are immutable.\n- Cloud Audit Logs entries are retained for 400 days.\n- Cloud Audit Logs feature is included in the price of Google Distributed Cloud software-only.\n- You can configure Google Distributed Cloud to write logs to disk or to Cloud Audit Logs.\n\nDisk-based audit logging\n\nIf Cloud Audit Logs is disabled explicitly, audit logs are written to a persistent\ndisk so that cluster restarts and upgrades don't cause the logs to disappear.\nGoogle Distributed Cloud software-only retains up to 1 GiB of audit log\nentries.\n\nAccess the disk-based audit logs by logging into control plane Nodes. The logs\nare located in the `/var/log/apiserver/` directory.\n\nCloud Audit Logs\n\nAdmin Activity audit log entries from all Kubernetes API servers are sent to\nGoogle Cloud, using the project and location that you specify when you\ncreate a user cluster. To buffer and write log entries to Cloud Audit Logs,\nGoogle Distributed Cloud deploys an `audit-proxy` daemon set that runs on the control\nplane nodes.\n\nLimitations\n\nOn bare metal, Cloud Audit Logs has the following limitations:\n\n- Data access logging isn't supported.\n- Modifying the Kubernetes audit policy is not supported.\n- Cloud Audit Logs isn't resilient to extended network outages. If the log entries can't be exported to Google Cloud, they are cached in a 10 GiB disk buffer. If that buffer fills, then the oldest entries are dropped.\n - One project can support up to approximately 1000 service accounts for use with Cloud Audit Logs. Multiple clusters can use the same service account.\n\nCreating a service account for Cloud Audit Logs\n\nBefore you can use Cloud Logging and Cloud Monitoring with\nGoogle Distributed Cloud software-only, you must first configure the following:\n\n1. Create a Cloud Monitoring Workspace within the Google Cloud project, if you\n don't have one already.\n\n In the Google Cloud console, click the following button and follow the\n workflow.\n\n [Go to Monitoring](https://console.cloud.google.com/monitoring)\n2. Click the following buttons to enable the required APIs:\n\n [Enable the Anthos Audit API](https://console.cloud.google.com/apis/library/anthosaudit.googleapis.com)\n\n [Enable the Stackdriver API](https://console.cloud.google.com/apis/library/stackdriver.googleapis.com)\n\n [Enable the Monitoring API](https://console.cloud.google.com/apis/library/monitoring.googleapis.com)\n\n [Enable the Logging API](https://console.cloud.google.com/apis/library/logging.googleapis.com)\n3. Assign the following IAM roles to the service account used by\n the Stackdriver agents:\n\n - `logging.logWriter`\n - `monitoring.metricWriter`\n - `stackdriver.resourceMetadata.writer`\n - `monitoring.dashboardEditor`\n\n| **Warning:** before deleting this service account, be sure to replace it with the new service account in the cluster configuration first! See [Rotate service\n| account keys](/kubernetes-engine/distributed-cloud/bare-metal/docs/how-to/update-secrets). If you forget to do this, follow [the guide to clean\n| up](/kubernetes-engine/distributed-cloud/bare-metal/docs/troubleshooting/observability#sa-leakage).\n\nAccessing Cloud Audit Logs \n\nConsole\n\n1. In the Google Cloud console, go to the **Logs Explorer** page in the\n **Logging** menu.\n\n [Go to the Logs Explorer](https://console.cloud.google.com/logs/query)\n\n If the **Legacy Logs Viewer** page opens, choose **Upgrade to the new\n Logs Explorer** from the **Upgrade** drop-down menu.\n2. Click **Query** to access the field for submitting queries.\n\n3. Fill the field with the following query:\n\n resource.type=\"k8s_cluster\"\n logName=\"projects/\u003cvar translate=\"no\"\u003ePROJECT_ID\u003c/var\u003e/logs/externalaudit.googleapis.com%2Factivity\"\n protoPayload.serviceName=\"anthosgke.googleapis.com\"\n\n Replace \u003cvar translate=\"no\"\u003ePROJECT_ID\u003c/var\u003e with your project ID.\n4. Click **Run query** to display all audit logs from clusters that were\n configured to sign in to this project.\n\ngcloud\n\nList the first two log entries in your project's Admin Activity log that\napply to the `k8s_cluster` resource type: \n\n gcloud logging read \\\n 'logName=\"projects/\u003cvar translate=\"no\"\u003ePROJECT_ID\u003c/var\u003e/logs/externalaudit.googleapis.com%2Factivity\" \\\n AND resource.type=\"k8s_cluster\" \\\n AND protoPayload.serviceName=\"anthosgke.googleapis.com\" ' \\\n --limit 2 \\\n --freshness 300d\n\nReplace \u003cvar translate=\"no\"\u003ePROJECT_ID\u003c/var\u003e with your project ID.\n\nThe output shows two log entries. Notice that for each log entry, the\n`logName` field has the value\n`projects/`\u003cvar translate=\"no\"\u003ePROJECT_ID\u003c/var\u003e`/logs/externalaudit.googleapis.com%2Factivity`\nand `protoPayload.serviceName` is equal to `anthosgke.googleapis.com`.\n\nAudit policy\n\nThe Kubernetes audit policy defines rules for which events are recorded as log\nentries and specifies what data the log entries should include. Changing this\npolicy to modify Cloud Audit Logs behavior isn't supported."]]
