FAQ: IAM Governance
The reason for this is because, in Azure, read permissions aren't logged. As a result, the last activity refers to the last management write activity that was performed.
Within the context of excessive permissions, Tenable One Cloud Exposure enforces a learning period for the following activities, during which no new findings (related to excessive permissions) are created:
- When a new human identity is created - the default is 90 days (configurable via IAM settings).
- When a new service/machine identity is created - the default is 90 days (configurable via IAM settings).
- When the permissions of an identity or resource are changed - 30 days
There is an automatic mechanism that filters and categorizes new services and permissions on an ongoing basis. Additionally, Cloud Exposure performs a periodic manual review of new services to verify that the automatic categorization is as accurate as possible.
How Tenable Detects Overprivileged Microsoft Entra ID Groups
Tenable evaluates whether a Microsoft Entra ID group has permissions that are not actively used by its members. Since groups themselves do not directly perform actions, the analysis is based on the permissions used by the group's members.
Technically, permissions are assigned to the group, but only the members use them. Cloud Exposure analyzes which permissions are actually used by the members to determine if any of the group’s assigned permissions are unnecessary.
It is based on each individual member. If any single member uses a specific permission assigned to the group, that permission is considered in use and is not flagged for removal.
Tenable aggregates the permissions used by all members. If Member A uses Permission X and Member B uses Permission Y, both X and Y are considered in use, and the group will be recommended to retain both.
Understanding the "Graph Too Large" Behavior
In certain cases, the permission data associated with a principal (such as a user or group) can grow beyond a threshold that is safe for the system to process. When this happens, we avoid attempting to load or render the data to prevent browser performance issues. As a result, both the graph and table views become unavailable.
Yes, there are two distinct scenarios:
- Scenario 1 — In rare cases, the data loads successfully but is too large to display as a graph. You can still view it in the table (list) format.
- Scenario 2 — In exceptional cases where the dataset size exceeds system limits, both the graph and table are withheld to avoid performance issues. A message is displayed instead.
In some cases, if the data is fetched successfully, narrowing the scope of the query - such as by selecting a smaller set of resources - may reduce the data size enough to allow it to be displayed.
Understanding Permission Discrepancies in Overprivileged Identity Findings
The summary only highlights fully unused services. If an identity uses at least one permission within a service (e.g., S3 or EC2), that service is excluded from the "excessive" count, even if it still contains many unused actions.
The system maps permissions based on functional impact. For example, certain EC2 permissions also grant the ability to manage ACM certificates. Therefore, ACM will appear in the graph even if the policy only explicitly mentions EC2.
Not exactly. The total count (e.g., 1,000 permissions) represents the entire scope of what the identity can do. This includes both necessary (used) and excessive (unused) permissions across all services.