The bug sat in the service for about four months, was fixed on the server side on 19 November, and needs no patching from customers.
· 4 min read
Microsoft has told customers that attackers exploited a flaw in its Entra Connect Health service to read administrator email addresses and tenant configuration information. BleepingComputer reported the breach on 21 November after administrators began sharing Microsoft's notification online. Microsoft says the exposed data was limited to contact and configuration metadata, and that passwords and authentication tokens were not involved.
What Microsoft says happened
Entra Connect Health is the monitoring service that helps organisations watch the health of their on-premises identity infrastructure, such as directory synchronisation. According to the notification shared by administrators, a vulnerability in the Connect Health API allowed unauthorised parties to read tenant-level metadata. Microsoft's notice says it detected the vulnerability on 9 November 2026 and that the flaw had been introduced on 23 July 2026. A server-side fix went in on 19 November, and by 20 November Microsoft says it had finished reviewing telemetry and notified the affected organisations.
Microsoft verified read-only access from its internal logs. It told BleepingComputer that it identified and addressed a security issue that could have resulted in unauthorised access to tenant-level contact and configuration metadata, and the notice states that no unauthorised configuration changes were found. The same notice says customers do not need to install updates or take other steps to close the hole, since the repair was made on Microsoft's side.
What was exposed, and what was not
The data returned by the API included email addresses for each tenant's technical contact and administrator accounts. It also included tenant IDs, display names, initial domain names, country and region information, Entra licence tiers, directory synchronisation status and Connect Health onboarding status. The API response also carried an expired certificate in a field called pksCertificate. Microsoft says that certificate poses no risk and is no longer used by the service, but that it had not been removed from Connect Health storage and was still included in the response.
Microsoft says the vulnerability did not expose user passwords, password hashes, authentication tokens, valid cryptographic keys, or details of on-premises servers and Active Directory topology. Administrators can run a Microsoft PowerShell script to see the categories of metadata that were exposed, although it shows current configuration values rather than records of what the attackers actually read.
The questions Microsoft has not answered
Microsoft has not said how many tenants were affected, who exploited the vulnerability, or when the attacks first began. It has also not explained why a flaw introduced in July took until November to detect. Some managed service providers reported receiving notices for several customer tenants, with one administrator saying the emails arrived hours apart, and more than a dozen administrators described their notifications in a public Reddit discussion. Only tenants that received a direct notice appear to have had confirmed unauthorised read access.
Why admin emails still matter
Metadata sounds harmless next to stolen passwords, but a list of administrator addresses is useful to attackers. The people named as technical contacts and global administrators are exactly the accounts a criminal would most like to phish, and a message that correctly references a company's Entra setup, domain and licence tier is far more convincing than a generic one. Security teams reading the notice have therefore focused on the follow-on risk rather than the breach itself.
Attackers are also automating the reconnaissance that precedes phishing and intrusion, as Amazon's report on AI agents finding a zero-day and gaining root access, published the same week, shows. Combining a harvested admin address with an automated pretext is a cheap next step.
What administrators should do
The practical advice from security researchers who have summarised the notice is short. Warn technical contacts and global administrators to expect targeted phishing that mentions Microsoft Entra and to treat unexpected messages with suspicion. Enforce phishing-resistant multi-factor authentication and Privileged Identity Management for privileged accounts, so a stolen address alone cannot lead to a takeover. Sign in directly to the Azure portal rather than through email links to check the notice under Azure Service Health, and review Entra audit logs for anything unusual. Organisations that did not receive a notice have no confirmed exposure, but it is reasonable to check the Service Health page.
A rough fortnight for Microsoft's cloud
The disclosure comes days after Microsoft's long Microsoft 365 outage on 12 and 13 November, which cut email for people around the world. Microsoft blamed that incident on a software defect introduced in July in its network gateway, combined with a failover and reduced server capacity, and said it found no sign of a cyberattack. The two incidents are unrelated according to what Microsoft has said, but both involve problems that sat in the system for months before they surfaced. For IT teams, the Entra notice is a reminder to keep administrator contact details current and to check that mail sent to those addresses reaches a person who will read it.
Published 22 November 2026. Gadgets365 will update this article as more information becomes available.
Editor
Pravin KankanePravin Kankane is the editor of Gadgets365, an independent technology publication based in Pune, India. He covers smartphones, AI, Apple, Android and consumer tech launches in India.


