Security Reports
Six reports on standing access across the estate and the changes that reshape your security posture. Filters, Subscribe, CSV, PDF and default views work as described in Using Reports.
10 Sections
Azure resource roles, not Entra ID directory roles
The access reports cover Azure resource role assignments. Microsoft Entra ID directory roles (such as Global Administrator) and directory changes (app registrations, conditional access) are not collected.
Prerequisites
- The View Resources permission and at least one completed scan. New & Revoked Privileged Access needs two.
- PIM Adoption needs Microsoft Entra ID P2 in the tenant.
Each report ends with a box describing what it can and can't show. The points below match those boxes.
Which Roles Count as Privileged
Standing Privileged Access, New & Revoked Privileged Access and PIM Adoption use the same definition.
- Critical
- Owner, Co-Administrator, Account Administrator, Service Administrator, User Access Administrator, Role Based Access Control Administrator.
- Management
- Contributor and any role ending in "Contributor", except data roles (names containing "Data", such as Storage Blob Data Contributor).
- Not privileged
- Everything else: Reader, data-plane roles and narrow administrator roles.
- Assignments are reported at management group, subscription, resource group and resource scope.
- Only direct assignments are shown. A group holding a privileged role counts as one identity; see its members in Role Assignments.
- These three reports filter by scope only.
The Change Window
Security Change Log, Weakened Security Settings and New & Revoked Privileged Access have a Window selector in the report body: 7d, 30d (default), 60d or 90d. A window of N days covers N UTC calendar days including today. When stored change history starts after the window begins, the report says when history begins, and longer windows show the same set.
Standing Privileged Access
Who can touch what, and should they: every Owner, User Access Administrator and Contributor at every scope, by identity, scope and how long the grant has stood. Uses the latest completed scan.
Controls: Filters (scope only).
- Privileged identities
- Count, split into users, groups and service principals.
- Critical-role holders
- Identities with a critical role.
- MG & subscription scope
- Privileged assignments at management group and subscription level.
- Longest standing
- Age of the oldest privileged grant. Amber at 180 days or more.
- Privileged identities table
- Top role, critical and management counts, scope reach (for example "2 MG · 1 sub · 3 RG · 4 res") and standing age. Critical identities first, up to 200.
- Non-human privileged access
- Service principals and managed identities with privileged roles, up to 100. Only shown when there are any.
- Standing only. PIM just-in-time assignments (activated eligibilities and time-bound active assignments) are excluded; the context line links to PIM Adoption with the count. Eligible-only assignments never appear. Without Entra ID P2, every assignment counts as standing.
- Standing age is measured from each assignment's creation date. Azure doesn't record who granted an assignment, so there is no "granted by" column.
CSV columns: Principal, Type, Highest category, Top role, Critical count, Management count, Standing days, Management group scopes, Subscription scopes, Resource group scopes, Resource scopes.
New & Revoked Privileged Access
Who gained elevated access lately: new privileged grants and revocations over the window, with the granting actor where the activity log can prove it.
Controls: Window (30d by default), Filters (scope only).
- New privileged grants
- Count, and how many have a known actor.
- New critical grants
- New Owner, User Access Administrator and administrator grants.
- Newly empowered identities
- Identities that gained privileged access.
- New power by scope
- Where new grants landed.
- Granted and revoked tables
- Identity, role, scope, date and the actor who granted or revoked it. Critical first, then newest; up to 200 of each.
- Compares the latest scan with the latest scan at or before the start of the window. With less history than the window, it compares with the earliest scan and says so.
- A grant is a principal, role and scope present now and not in the baseline scan. Grants made and removed between two scans don't show.
- Granted by and Revoked by come from the Azure activity log, kept about 365 days.
CSV columns: Change (Granted/Revoked), Principal, Type, Role, Category, Scope, Actor, Actor timestamp.
A weekly "who gained admin" email
Set the window to 7d and Subscribe weekly to get a recurring email of new privileged access.
PIM Adoption
How access is granted, just-in-time or standing: the share of role assignments and principals managed through PIM eligibility versus permanent assignment, and who to convert next.
Controls: Filters (scope only).
- PIM adoption (privileged)
- Share of privileged grants managed through PIM, plus the share across all roles.
- Standing privileged grants
- Permanent Owner, User Access Administrator and Contributor-family grants.
- PIM-eligible grants
- Just-in-time grants, all roles.
- Active via PIM now
- Activated or time-bound grants at scan time, all roles.
- How access is granted
- PIM eligible, PIM active and permanent, for all roles and for privileged roles only.
- Principals by access model
- Users and groups that are PIM-only, mixed or standing-only.
- PIM conversion candidates
- Identities still holding standing privileged grants, worst first, with their PIM coverage. Up to 200.
- Needs Microsoft Entra ID P2. Without it, the report says P2 wasn't detected instead of reporting 0% adoption, and CSV is disabled.
- If PIM data couldn't be collected for some scopes, the context line says so and counts may be low.
- Service principals and managed identities are excluded, because PIM activation is an interactive sign-in workflow. Their access is under Non-human privileged access in Standing Privileged Access.
- A currently activated eligibility counts once, as active. A PIM assignment with no end date counts as standing.
CSV columns: Principal, Type, Top privileged role, Category, Standing privileged grants, PIM-eligible (privileged), PIM-active (privileged), PIM coverage %.
Security Change Log
Changes to the resources that govern access and security posture, attributed to a named actor.
Controls: Window (30d by default), Filters (all dimensions).
- Security changes
- Created, modified and deleted.
- Controls deleted
- Security resources removed.
- Who's changing security resources
- Created, modified, deleted and total for the top 15 actors.
- Security change log
- Resource, area, change, actor and time. Newest first, up to 100.
Areas
- Network security
- NSGs, Azure Firewalls and firewall policies, public IPs, DDoS plans, Bastion, Application Gateways, Front Door, WAF policies, private endpoints, route tables.
- Identity & access
- Role assignments, role definitions, managed identities.
- Key & secret management
- Key Vault.
- Policy & governance
- Policy assignments and definitions, Microsoft Defender for Cloud settings, diagnostic settings.
- Data stores (storage, databases) aren't included; Weakened Security Settings and Public Exposure cover their risky settings.
- It flags changes by resource type and doesn't judge whether a change made things weaker. Everyday resource changes are in Infrastructure Change.
CSV columns: Resource, Type, Category, Change type, Change time, Actor, Actor kind, Subscription, Resource group.
Weakened Security Settings
Configuration changes that moved a security setting to a weaker value, attributed to a named actor.
Controls: Window (30d by default), Filters (all dimensions).
The cards count weakenings, critical weakenings and resources affected, and show how many changes have a named actor. A chart groups them as Public access, Network exposure, Transport security or Data protection, and the table lists each one with before → after, severity, actor and time, most severe first (up to 100).
What it checks
- Minimum TLS version lowered
- Storage, SQL server, App Service: High
- HTTPS-only / secure transfer turned off
- Storage, App Service: High
- Public network access turned on
- Storage, SQL, Cosmos DB, Key Vault: Critical
- Public network access turned on
- App Service: High
- Blob anonymous access turned on
- Storage: High
- Firewall default changed from Deny to Allow
- Key Vault: High
- Soft-delete turned off
- Key Vault: High
- Purge protection turned off
- Key Vault: Critical
- Infrastructure (double) encryption turned off
- Storage: Medium
- Public IP newly attached
- Public IP: Medium
- NSG rule source changed to the internet
0.0.0.0/0,*,::/0, Internet or Any: Critical if confirmed as an inbound Allow, High if the change doesn't show direction and access.
- Only changes to existing resources. A resource created insecure shows in Public Exposure.
- Both the before and after values must be recorded; nothing is assumed.
- It is a focused list, not a full security audit.
CSV columns: Resource, Type, Category, Severity, Setting, Before value, After value, Reason, Change time, Actor, Subscription, Resource group.
Public Exposure
A point-in-time inventory of everything reachable from the internet, with what is new since the last scan. Uses the latest completed scan.
Controls: Filters (all dimensions).
The cards count resources with findings, exposed data stores, wide-open NSG rules and findings new since the last scan. Tiles break exposure down by type with a "+N new" count, and the findings table marks new rows with a NEW badge (up to 300).
What it flags
- Public IP
- Attached to a NIC, load balancer, application gateway, firewall, Bastion, VPN gateway or NAT gateway: High. Unattached public IPs are in Orphaned Resources.
- NSG
- Inbound Allow from the internet: Critical when the ports include 22, 3389, 135, 445, 23, 5985, 5986, 1433, 3306, 5432, 6379, 27017, 9200, 5984, 11211 or any/all, otherwise High.
- Storage
- Blob public access on, or public network access enabled: High. Weak TLS is noted.
- SQL server
- Public network access enabled: High.
- Key Vault
- Public network access enabled, or firewall default Allow: High.
- App Service
- Public network access enabled: Medium. HTTPS-only off is noted.
- Positive evidence only: a missing setting is never assumed open.
- A private endpoint doesn't turn off public network access, so an enabled flag is reported regardless.
- "New" compares with the previous completed scan. The first scan has nothing to compare with and is point-in-time only.
CSV columns: Resource, Type, Category, Severity, Reason, Detail, New (Yes/No), Subscription, Resource group, Location.
Badges
- Principal type
- User, Group or Service principal.
- Actor kind
- Person is a user; Automation is an application or managed identity; System is Azure itself or an unrecognized client.
- Unknown actor
- No actor could be proven. StratoLens never guesses.
- Severity
- Critical, High or Medium. Critical rows have a red marker.
Troubleshooting
New & Revoked Privileged Access is empty after install
Answer
It needs two completed scans to compare. Run another scan and the report shows what changed in between.
PIM Adoption says P2 not detected
Answer
The tenant has no Microsoft Entra ID P2 license, so there is no PIM data to measure.
Granted by says Unknown
Answer
The activity log didn't record who made the grant within its retention. StratoLens doesn't guess.
A 90-day window shows the same as 30 days
Answer
Change history only goes back to the date in the "change history begins" hint. Azure keeps about 14 days of change history before StratoLens starts collecting.
My tag filter doesn't change Standing Privileged Access
Answer
Standing Privileged Access, New & Revoked Privileged Access and PIM Adoption use scope filters only, because role assignments aren't tagged resources.