Microsoft's Connect Health API Sat Open From July to November and Attackers Read Admin Emails
Microsoft fixed an Entra Connect Health API bug on November 19. Attackers had read-only access to admin emails and tenant metadata. No customer patch is required.
Microsoft spent four months with a hole in the API that watches on-premises identity, then told customers the attackers had already read their admin inboxes' addresses.
Lawrence Abrams reported on November 21 that Microsoft discovered a vulnerability in the Entra Connect Health API on November 9. The bug had been introduced on July 23. A server-side fix went out on November 19. The investigation wrapped on November 20, and Azure Service Health notices started going to affected tenants.
Connect Health is the cloud console for the boxes that sync Active Directory into Entra. The flaw let unauthorized callers read tenant and service-monitoring metadata, and Microsoft says it could also have let them change monitoring configurations. In the tenants it checked, it found only reads. "We did not identify any unauthorized configuration changes," the notice said.
What left the building
Microsoft says the exposed fields included email addresses for tenant technical contacts and administrator accounts; tenant IDs, display names, initial domain names, country codes, and region; Entra license tiers; whether directory sync was on and when it last ran; Connect Health onboarding status; and an expired certificate once used to register Connect Health agents with an internal Microsoft service retired in late 2024. The company says that certificate was no longer in use, but the API still handed it back.
It also listed what did not leak: user passwords, password hashes, authentication tokens, valid cryptographic keys, on-premises server and Active Directory topology, and Azure subscription details.
A Microsoft spokesperson told BleepingComputer: "We identified and addressed a security issue affecting Microsoft Entra Connect Health that could have resulted in unauthorized access to tenant-level contact and configuration metadata. We deployed a fix on November 19, 2026, and have notified impacted customers through Azure Service Health, providing guidance and support." Microsoft would not add a tenant count, an attacker, or a start date for the reads. More than a dozen admins had posted the notice by the time Abrams wrote. Some MSPs said they got copies for several customer tenants, hours apart.
No patch for you, plenty of phishing bait
Microsoft says customers do not need to install anything. The fix is on its side. It wants security teams watching for phishing and social engineering aimed at those admin addresses, phishing-resistant MFA on privileged accounts, a pass over Connect Health settings, and a look at Entra audit and sign-in logs. A read-only PowerShell script at aka.ms/ConnectHealth/scripts shows the kinds of metadata the API holds now. It does not reconstruct what an attacker actually pulled.
Admin mailboxes are already a favorite door. Cyberpresso covered a summer campaign that replayed stolen Microsoft 365 sessions past MFA and planted inbox rules. A list of global-admin emails, tenant IDs, and sync timestamps is the targeting pack for the next one.
Microsoft has not published a CVE in the notices described so far. The window it did publish is long enough: July 23 to November 19, with the company noticing on November 9.
Some offers on this page may be paid placements or contain affiliate links.
Cyberpresso: daily cyber & AI brief
Free daily newsletter, read in 5 minutes.
Subscribe free