Microsoft Entra flaw exploited to steal tenant data

Microsoft has told organisations that attackers exploited a flaw in its Entra Connect Health service to read tenant data and administrator email addresses. The company disclosed the incident in a security notification sent to administrators on 20 November, after an investigation concluded that unauthorised read-only access to tenant-level information had taken place.
Entra Connect Health is a cloud monitoring service that lets administrators watch the health, performance and alerts of their on-premises identity infrastructure through the Microsoft Entra admin centre. The flaw sat in the service's application programming interface, known as an API, which other software uses to exchange data with the service.
What was exposed
According to Microsoft, the information exposed through the flaw included email addresses belonging to tenant technical contacts and administrator accounts. The API responses also revealed tenant identifiers, display names, initial domain names, country codes and regional information, along with Microsoft Entra licence tiers and directory synchronisation status, including whether synchronisation was enabled and the time of the last synchronisation.
The responses also included Connect Health onboarding status and an expired certificate that had previously been used to register Connect Health agents with a Microsoft internal service decommissioned in late 2024. Microsoft said the certificate was no longer in use but had remained in Connect Health storage and was still being returned by the API. The company stressed that the expired certificate poses no risk.
How long the flaw was live
Microsoft said the vulnerability was introduced into the Connect Health API on 23 July 2026 and was discovered on 9 November 2026, meaning it was live for more than three months before it was found. The company deployed a server-side fix on 19 November to block further exploitation, finished its investigation on 20 November and began notifying affected organisations the same day through Azure Service Health.
Because the fix was applied on Microsoft's own servers, customers do not need to install any update. Microsoft also published a PowerShell script that lets administrators see the types of tenant metadata that could have been exposed, although the script shows current configuration values rather than the information that was actually accessed during the incident.
What was not exposed
Microsoft said the flaw did not expose user passwords, password hashes, authentication tokens, valid cryptographic keys, on-premises server and Active Directory topology details, or Azure subscription information. Investigators also found no evidence of unauthorised configuration changes, even though the flaw could in theory have allowed monitoring configurations to be modified.
What remains unknown is who exploited the vulnerability, how many tenants were affected and when the attacks first began. Microsoft has not disclosed any of those details, and more than a dozen administrators have reported receiving the notification in a Reddit thread, with some managed service providers saying they received notices for multiple customer tenants.
What administrators should do now
Microsoft recommends that organisations alert their security teams to the risk of targeted phishing and social engineering attacks aimed at the exposed administrator email addresses and tenant information. Administrators should make sure privileged accounts use phishing-resistant multifactor authentication, review their Connect Health configurations and examine Microsoft Entra audit and sign-in logs for suspicious activity.
No CVE identifier has been assigned to the flaw, which was fixed entirely on Microsoft's side. The incident is a reminder that even a monitoring service can become an entry point: the data taken was configuration and contact information rather than credentials, but in the hands of an attacker it can make a convincing phishing message much easier to build.