Resolve Cloud Onboarding Permission and Role Errors (AWS, Azure, GCP)

Objective
Most cloud onboarding failures come down to permissions: the role or service account that Tenable One Cloud Exposure uses is missing an action, is blocked by an organization policy, or was created with the wrong scope. Use this playbook when an account or organization does not finish onboarding, shows a Connected with Issues or Failed status, or returns a Missing permissions error even though the setup script appeared to complete successfully.

Note: A setup script can finish without errors and still leave the connection short of permissions (e.g., when it runs under an identity that itself lacks the rights to grant everything Tenable needs). If the Finish step reports missing permissions, a script that completed without errors does not confirm complete permissions.

Step 1: Validate the Role and Service Account Setup

Note: These checks apply regardless of CSP.

To confirm the role or service account is correct:

  1. Deploy the role or service account using the provided onboarding script or Terraform module. A manually trimmed permission set is the most common cause of a Missing permissions error.
  2. Onboard the entire organization rather than individual accounts. Connecting at the organization level avoids onboarding accounts one at a time and gives Cloud Exposure ongoing visibility as new accounts and their resources are added. Run the setup at the organization so member accounts are discovered and connected automatically.
  3. After you change permissions, confirm the fix using the following methods:
    • Reopen the onboarding wizard and click Finish.
    • Wait for the next orchestration cycle, then check whether the issue cleared.

Step 2: Check Cloud-Specific Blockers

AWS

  • Service Control Policies (SCPs) — An org-level SCP can silently block the Cloud Exposure role even when the role and CloudFormation stack look correct. Confirm that no SCP denies the actions Cloud Exposure needs. This is a frequent cause of member accounts stuck in a Connected with Issues status.
  • StackSets across the organization — For an AWS Organization, confirm that the CloudFormation StackSet deployed the role into every member account. Accounts without the role show a Connected with Issues or Failed status even when the management account looks fine.
  • Cross-account CloudTrail and KMS — If your CloudTrail bucket is in a separate log-archive account, or encrypted with an AWS Key Management Service (KMS) key from another account, the Cloud Exposure role needs cross-account read access and KMS decrypt permissions.

Azure

  • Built-in roles vs. custom roles. Cloud Exposure checks for specific built-in role assignments. A custom role that looks equivalent can still fail the check. Assign the built-in roles that Tenable specifies.
  • Directory permissions — Onboarding a Microsoft Entra ID tenant or organization requires sufficient directory rights (e.g., a Global Administrator role to grant consent). A Missing permissions to retrieve organization data error in the final step usually points to this cause.
  • Second tenant or single sign-on (SSO) authentication loop. If your primary container enforces corporate-domain SSO, onboarding a second Microsoft Entra ID tenant can cause an authentication loop. To resolve this, use a private or incognito browser window with the target tenant's administrator credentials, or register the application manually as a service principal and enter the tenant ID, client ID, and secret directly.

GCP

  • Organization scope vs. project scope. When you onboard individual projects, the Project Owner role alone is not sufficient. Assign the required role at the organization level.

Step 3: Verify the Connection

To verify the connection status:

  1. Reopen the onboarding wizard and click Finish to validate the key permissions. This checks the key permissions only, not the full set.
  2. Wait for the next orchestration cycle.
  3. Confirm that each member account, subscription, or project shows a Connected status, not just the management account, and that data begins to populate.
  4. Allow the first full analysis to complete. For a large organization, findings can take several hours to populate.
Important: If the account still does not connect, or accounts remain stuck after the next orchestration cycle despite a correct role, SCP, and StackSet, contact Tenable Support. Provide your container ID, the cloud provider, and the exact error text.