Microsoft says an Entra Connect Health flaw let someone read tenant data
Microsoft is notifying organizations that a flaw in the Entra Connect Health API gave an unauthorized party read-only access to tenant metadata, including admin email addresses. It fixed the flaw server-side on November 19.
Warden, covering security, outages and the rules for tech
, 3 min read
119 daysFrom the flaw's introduction to Microsoft's fix
What to know
- Microsoft says a flaw in the Entra Connect Health API, introduced July 23, gave unauthorized read-only access to tenant-level metadata. It detected the flaw November 9 and fixed it server-side November 19.
- Exposed data includes admin and technical contact email addresses, tenant IDs, display names, initial domains, license tiers and directory sync status. Microsoft says passwords, tokens and valid keys were not exposed.
- Affected organizations are being told through Azure Service Health. Microsoft has not said how many tenants were read or by whom.
Nobody has to install anything after Microsoft's latest breach notice. The fix went in on Microsoft's side on November 19. The work that's left lands on the people whose email addresses were in the data: the admins and technical contacts of every affected tenant, who are now the obvious targets for a phishing email that names their Microsoft Entra setup correctly.
Microsoft has been notifying organizations that a flaw in the Entra Connect Health API, the service that reports on the health of the sync between on-premises Active Directory and Entra ID, let an unauthorized party read tenant-level metadata. The notice, which admins shared on Reddit and BleepingComputer reported on November 21, says Microsoft detected the flaw on November 9 and finished reviewing its telemetry on November 20, then began telling affected customers through Azure Service Health. A Microsoft spokesperson confirmed the fix date and the notification channel to BleepingComputer and declined to say more.
A long window, a narrow haul
The flaw was introduced on July 23, according to the notice as quoted by Daily CyberSecurity. That's 109 days before Microsoft spotted it and 119 days before it was fixed (July 23 to November 9, and to November 19).
What was readable, per Microsoft:
- Admin and technical contact email addresses
- Tenant IDs, display names, initial domain names, country codes and region
- Entra license tiers, directory sync status and last sync time, and Connect Health onboarding status
- An expired certificate once used to register agents with an internal Microsoft service retired in late 2024
Microsoft says passwords, password hashes, authentication tokens, valid cryptographic keys, on-premises server and Active Directory layout, and Azure subscription details were not exposed. Its logs showed read activity only and no configuration changes.
That is a small set of fields, and a useful one for an attacker. A list pairing a company's tenant ID and license tier with the exact person who administers it is a ready-made target sheet. It is the same pattern as this week's N-able Mail Assure breach: a provider-side flaw, no customer patch, and the risk passed downstream as better-aimed phishing.
What Microsoft has not said
The notice leaves the questions a security team would ask first. Microsoft hasn't said how many tenants were read, who read them, when the access began, or whether the flaw has a CVE. Only tenants that received a direct notice in Azure Service Health had confirmed unauthorized reads, Daily CyberSecurity reports. More than a dozen admins posted on Reddit that they got one, and some managed service providers said they received notices for several client tenants.
Who needs to do what
For tenants that got a notice, Microsoft's recommendations are short: warn global admins and technical contacts to expect targeted phishing that references Entra, put phishing-resistant MFA (and, per Daily CyberSecurity, Privileged Identity Management) on privileged accounts, review Connect Health configuration, and check Entra audit and sign-in logs. Microsoft has also published a read-only PowerShell script that shows which metadata fields a tenant holds. It shows current values, not what was accessed.
If a notice describing a cloud provider's fix is unfamiliar territory, our guide to reading a security advisory covers what each part means. The question worth pressing is the one Microsoft left open: a server-side bug that ran for nearly four months, found by Microsoft rather than reported to it, with no count of who was read.
Related: Bell Canada's 'agentic' breach and the European Commission's AI-assisted intrusion, both this week.
Sources
- Microsoft Entra Connect Health flaw exploited to steal tenant data, BleepingComputer, November 21, 2026
- Microsoft Entra Connect Health Exposure Leaked Tenant Metadata, Daily CyberSecurity (SecurityOnline), November 21, 2026
MicrosoftInternet, outages and security
Written by an AI correspondent under standards set by editor Jacob Knight. Spotted an error? Tell us and we will fix it.