Find the insights and best practices about our product.
CloudSensor Overview

Introduction

CloudSensor is a cloud security monitoring service that connects to Azure tenants and AWS accounts to discover security alerts, monitor unauthorized changes, and assess role-based access controls. Every connection scans automatically once a night, and you can also run an on-demand scan for a single connection at any time.

Managing CloudSensor Integrations

Navigate to CloudSensor Integrations

In the left navigation, select Integrations, then select the CloudSensor Integrations tab.

CloudSensor Integrations List

The CloudSensor Integrations table lists every cloud connection in your tenant. Use it to check that each connection is healthy and scanning. The table shows these columns:

  • Name: The connection's name. Select it to open the connection.
  • Health Status: The current state of the connection. A connection that has not been activated shows New.
  • Description: The description entered when the connection was created.
  • Provider: The cloud provider, Azure or AWS.
  • Account ID: The Azure Tenant ID or the 12-digit AWS Account ID.
  • Last Updated: When the connection was last scanned. A connection shows Never until its first scan.
  • Registration: The connection's activation state. A new Azure connection shows CLICK TO ACTIVATE. AWS connections show ACCESS VERIFIED, ACCESS LOST, ACCESS FAILED or AWAITING VERIFICATION.
  • Scan: The Run scan button, which starts an on-demand scan for the connection.

The three-dot action menu at the end of each row holds the management actions described below.

Managing Existing Integrations

Each connection provides management options through the three-dot action menu:

  • Edit Integration: Change the connection's name and description.
  • Verify Access: Read from the cloud account and report what Cranium can read. See Checking Access below for each result.
  • Delete Integration: Remove the connection and stop future scans.

The cloud provider and the Azure Tenant ID or AWS Account ID cannot be changed after the connection is created. To use a different account, create a new connection.

Edit Integration and Verify Access require the CloudSensor_Integrations_Update permission. Users with Read only do not see Verify Access. By default only the Admin role has Update. Delete Integration requires the CloudSensor_Integrations_Delete permission. If you have permission for none of the actions in the menu, the three-dot action menu is hidden.

Each discovery run keeps Health Status up to date. Running Verify Access on an existing Azure connection also updates its status. A connection that fails the check moves out of ACTIVATED to show that access has been lost or failed.

When an Azure grant is narrowed, the connection shows a permission-change alert instead of marking the assets in those scopes as decommissioned. Accept the current access to clear the alert.

Use Run scan to refresh a single connection's findings without waiting for the nightly scan, for example right after a change in your cloud account that you want reflected quickly. Only one discovery job can run per tenant at a time. If a scan is already running, whether the nightly scan or another connection's on-demand scan, starting a new one returns a conflict message instead of starting a second job.

Connecting a Cloud Account

Prerequisites

For every connection:

  • An active Cloud Sensor license for your organization
  • The CloudSensor_Integrations_Create permission on your role

For an Azure tenant:

  • The Azure Tenant ID of the environment to monitor
  • An administrator of the Azure directory who can consent and grant read access

For an AWS account:

  • The 12-digit AWS Account ID
  • A person in the AWS account with the IAM permissions needed to deploy the read-only role. The Set up AWS access dialog lists these permissions.

Create Integration

Select Add New CloudSensor Integration to open the Create CloudSensor Integration window. The window lists its four steps on the left: Details, Consent, Grant access and Verify access. Select Continue to move to the next step, Previous to go back, or Cancel to close the window.

  1. Details: Under CloudSensor Details, enter a Name and a Description. Both are required. Choose the Cloud Provider, then enter the Azure Tenant ID or the AWS Account ID. The provider cannot be changed after the connection is created.
  2. Consent (Azure only): An administrator of your Azure directory consents to Cranium signing in. Select Open Azure consent to start. Consent creates Cranium's service principal in your directory. It grants no access to your resources on its own. The next step gives that principal read access.
  3. Grant access: For Azure, this step is Grant read access. For AWS, this step is Deploy the read-only role.
  4. Verify access: Select Check access to test the connection. If a result needs a fix, make the change and select Check again. Finish is enabled once access is verified. Skip for now creates the connection without verifying.

For Azure, read access lets Cranium read security alerts and configurations, monitor role assignments, read compliance and audit logs, and detect unauthorized changes.

Connecting an AWS Account

Finish stays disabled until access is verified, and an AWS account can only be verified after the read-only role exists in that account. The role is deployed from the connection's row, so the connection must exist first. To connect an AWS account:

  1. Complete the Details step, then select Skip for now in the Verify access step to create the connection.
  2. Select set up AWS access on the connection's row and deploy the read-only role. See Setting Up AWS Access below.
  3. Select verify access on the connection's row to check the role.

Setting Up AWS Access

The read-only role gives Cranium access to discover AI resources in your AWS account. Select set up AWS access on the connection's row to open Set up AWS access for the connection. The dialog shows:

  • External ID and Cranium's AWS account ID: The values the role trusts. The deployment templates already contain them.
  • Required IAM permissions: The permissions the person deploying the role needs in your AWS account.
  • Region selector: The AWS region to deploy the role from.
  • Deploy to AWS: Opens a one-click CloudFormation deployment in your AWS account.
  • Download the CloudFormation template: Downloads the template to deploy yourself.
  • Terraform template and read-only policy: The Terraform template and the policy the role receives, for review.

Cranium works out the role ARN from the account ID, so you do not need to paste it. An override is available for AWS GovCloud and China accounts.

Cranium discovers the regions to scan from each account. A region blocked by policy does not fail the whole scan.

Checking Access

Verify Access in the three-dot menu, verify access on the connection's row, and Check access in the Verify access step all perform a live read in your cloud account. Each check reports one of these results:

  • Access verified: Cranium can read the account.
  • Access verified — no AI resources found: Cranium can read the account but found no AI resources in it.
  • Access is missing: The role or grant does not allow every action Cranium needs. For AWS, the missing actions are listed under Grant these actions to the role: with a copy button. Add them to the role and check again.
  • Not connected: Cranium has no working connection to the account.
  • Access check failed: The check could not complete.
  • Access check not available: The check cannot run for this connection.

The check reports missing access even when an AWS Organizations service control policy blocks discovery. A service control policy is an account-wide AWS rule that can block actions for every role in the account.

Troubleshooting Azure Access

If an Azure connection does not reach Access verified:

  • Verify that the Azure administrator privileges are active.
  • Confirm that the Azure Tenant ID is correct. To use a different tenant, create a new connection.
  • Check network connectivity and firewall settings.
  • Validate your Cloud Sensor license status.

After access is verified, CloudSensor begins discovery. The connection's Last Updated value changes from Never after its first scan.

Viewing Discovered Assets

The AI models and resources that your connections discover appear on the Cloud Assets page. In the left navigation, select Artifacts, then select Cloud Assets.

The Connection column shows the connection that discovered each asset, and the Connection filter shows only the assets from one connection. When you open an asset, its details also show the connection. If a connection has been deleted, its assets show Unavailable in the Connection column.

If no connection exists yet, the page shows No cloud connections configured with a Go to Integrations link.

Did this answer your question?