如果您使用 VPC Service Controls,只有在授予 Security Command Center 服務帳戶存取權時,才能探索及更新 Security Command Center 中的資產。
如要啟用資產探索功能,請授予 Security Command Center 服務帳戶存取權。服務帳戶即可完成資產探索,並在 Google Cloud 控制台中顯示資產。服務帳戶名稱的格式為 service-org-organization-id@security-center-api.iam.gserviceaccount.com。
查看、編輯、建立及更新調查結果和資產
您可以在機構、資料夾或專案層級授予 Security Command Center 的 IAM 角色。您能否查看、編輯、建立或更新發現項目、資產和安全性來源,取決於您獲准的存取層級。如要進一步瞭解 Security Command Center 角色,請參閱存取權控管。
未收到通知或通知延遲
在某些情況下,通知可能會遺漏、捨棄或延遲:
可能沒有任何發現項目符合 NotificationConfig 中的篩選條件。如要測試通知,請使用 Security Command Center API 建立發現項目。
Security Command Center 服務帳戶必須在 Pub/Sub 主題中具備 securitycenter.notificationServiceAgent 角色。服務帳戶名稱的格式為 service-organization-id@gcp-sa-scc-notification.iam.gserviceaccount.com。
如果網址對應至同機構中其他專案預留的 IP 位址,就會發生這項錯誤。如要解決這個問題,請在目標專案中,定義以專案 VM 或 HTTP 負載平衡器為目標的安全掃描作業。
網址對應至多個 IP 位址。
請確認對應至此網址的所有 IP 位址皆已保留給同個專案。如有至少一個 IP 位址未保留給相同專案,掃描建立或編輯/更新作業即會失敗。
Model Armor
本節提供疑難排解步驟,協助您解決使用 Model Armor 時遇到的問題。
對 Model Armor 的所有 API 呼叫都會傳回 404 Not Found 錯誤
建立 Private Service Connect,連線至 Model Armor API。如果使用 Private Google Access 存取 Model Armor 區域端點 (REP),或未建立 Private Service Connect,通常會發生這項錯誤。詳情請參閱「透過 Private Service Connect 端點存取區域端點」。
Sensitive Data Protection 篩選器發生錯誤或遭到略過
請確認下列事項:
Sensitive Data Protection 範本與所呼叫的 Model Armor 端點位於相同區域。
提出 Model Armor 要求的服務代理在含有 Sensitive Data Protection 範本的專案中,具有 dlp.User 和 dlp.Reader 角色。
發生這項錯誤的原因是 SanitizeUserPrompt 或 SanitizeModelResponse 要求發生用戶端錯誤,或是 Sensitive Data Protection 範本有問題。
Model Armor 全球端點錯誤
請確認您是向適當的區域端點發出 API 要求,而非全域端點。
Model Armor 僅支援對區域端點執行下列作業:
對範本執行建立、讀取、更新、刪除及列出作業。
SanitizeUserPrompt 和 SanitizeModelResponse API 要求。
如果您向全域端點提出這些作業的 API 要求,就會看到下列錯誤。
{
"error": {
"code": 403,
"message": "Write access to project '<PROJECT_ID>' was denied",
"status": "PERMISSION_DENIED"
}
}
[[["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:2025-09-09 (世界標準時間)。"],[],[],null,["| Standard, Premium, and Enterprise [service tiers](/security-command-center/docs/service-tiers)\n\nLearn about troubleshooting steps that might be helpful if you experience the\nfollowing problems while using Security Command Center.\n| **Note:** For more troubleshooting information, see [Error messages](/security-command-center/docs/error-messages). In addition, Security Command Center provides error detectors, which report configuration errors that prevent Security Command Center and its services from working properly. Each error finding includes suggested remediation steps. For more information, see [Security Command Center errors](/security-command-center/docs/concepts-scc-errors).\n\nEnabling Security Command Center fails\n\nEnabling Security Command Center most commonly fails if your organization policies\n[restrict identities by domain](/resource-manager/docs/organization-policy/restricting-domains). You and your\nservice account must be part of an allowed domain:\n\n- Make sure you sign in to an account that's in an allowed domain before you try to enable Security Command Center.\n- If you're using an `@*.gserviceaccount.com` service account, add the service account as an identity in a group within an allowed domain.\n\nAssets in Security Command Center aren't updating\n\nIf you're using VPC Service Controls, assets in Security Command Center can only be\ndiscovered and updated when you grant access to the Security Command Center service\naccount.\n\nTo enable asset discovery, [grant access](/security-command-center/docs/how-to-use-security-health-analytics#service-perimeter) to the Security Command Center service account. This allows the service account\nto complete asset discovery and display assets in the Google Cloud console.\nThe service account name is in the form of\n`service-org-`\u003cvar translate=\"no\"\u003eorganization-id\u003c/var\u003e`@security-center-api.iam.gserviceaccount.com`.\n\nViewing, editing, creating, and updating findings and assets\n\n\nThe IAM roles for Security Command Center can be granted at the organization,\nfolder, or project level. Your ability to view, edit, create, or update findings, assets,\nand security sources depends on the level for which you are granted access. To learn more about\nSecurity Command Center roles, see [Access control](/security-command-center/docs/access-control).\n\nMissing or delayed notifications\n\nIn some situations, notifications might be missing, dropped, or delayed:\n\n- There might not be any findings that match the filters in your `NotificationConfig`. To test notifications, use the Security Command Center API to [create a finding](/security-command-center/docs/how-to-api-create-manage-findings#creating_a_finding).\n- The Security Command Center service account must have the `securitycenter.notificationServiceAgent` role on the Pub/Sub topic. The service account name is in the form of `service-`\u003cvar translate=\"no\"\u003eorganization-id\u003c/var\u003e`@gcp-sa-scc-notification.iam.gserviceaccount.com`.\n - If you remove the role, notification publishing is disabled.\n - If you remove the role and then grant the role again, notifications are delayed.\n- If you delete and recreate the Pub/Sub topic, notifications will be dropped.\n\nWeb Security Scanner\n\nThis section contains troubleshooting steps that you might find helpful if you\nhave problems using Web Security Scanner\n\nScan errors for Compute Engine and GKE\n\nIf the URL for a scan is misconfigured, Web Security Scanner rejects it. Possible\nreasons for rejection include:\n\nURL has an IP address that is ephemeral\n\nMark this IP address as static:\n\n- For an application on a single VM, reserve the IP address on the VM\n- For an application behind a load balancer, reserve the IP address on the load balancer.\n\nURL is mapped to a wrong IP address\n\nTo fix this finding, refer to the instructions from your DNS registrar service.\n\nURL is mapped to an ephemeral IP address of the same VM\n\nMark this IP address as static.\n\nURL is mapped to a reserved IP address\n\nThis error happens when the URL is mapped to an IP address that's reserved in a\ndifferent project of the same organization. To resolve this, define security\nscans for the VM or HTTP load balancer in the project for which it is defined.\n\nURL is mapped to more than one IP address.\n\nMake sure that all IP addresses that are mapped to this URL are reserved for the\nsame project. If there is at least one IP address that isn't reserved for the\nsame project, the Scan Create or Edit or Update operation fails.\n\nModel Armor\n\nThis section contains troubleshooting steps that you might find helpful if you\nhave problems using Model Armor.\n\nAll API calls to Model Armor return a 404 Not Found error\n\nEstablish a Private Service Connect to the\nModel Armor APIs. This error typically occurs when Model Armor\nregional endpoints (REPs) are accessed using Private Google Access or without\na Private Service Connect. For more information, see\n[About accessing regional endpoints through Private Service Connect endpoints](/vpc/docs/about-accessing-regional-google-apis-endpoints).\n\nSensitive Data Protection filter gives an error or is skipped\n\nVerify the following:\n\n- The Sensitive Data Protection template is in the same region as the Model Armor endpoint being called.\n- The service agent making the Model Armor request has the `dlp.User` and `dlp.Reader` roles in the project containing the Sensitive Data Protection template.\n\nThis error occurs due to client errors in the `SanitizeUserPrompt` or `SanitizeModelResponse`\nrequests or issues with the Sensitive Data Protection template.\n\nModel Armor global endpoint errors\n\nEnsure that you are making the API requests to the appropriate\nregional endpoint instead of the global endpoint.\n\nModel Armor supports the following operations only\non its regional endpoints:\n\n- Create, read, update, delete, and list operations on templates.\n- `SanitizeUserPrompt` and `SanitizeModelResponse` API requests.\n\nIf you make the API requests for those operations to the global endpoint, you\nwill see the following error. \n\n {\n \"error\": {\n \"code\": 403,\n \"message\": \"Write access to project '\u003cPROJECT_ID\u003e' was denied\",\n \"status\": \"PERMISSION_DENIED\"\n }\n }\n\nWhat's next\n\nLearn about [Security Command Center errors](/security-command-center/docs/concepts-scc-errors)."]]