Microsoft Entra Connect Health API flaw exploited: admin email addresses and tenant metadata exposed, Microsoft notifying affected tenants
SockAnalyst · Security+ certified analyst.
What happened
Microsoft is notifying organizations through Azure Service Health that an exploited vulnerability in the Microsoft Entra Connect Health API gave unauthorized parties read-only access to tenant-level metadata. According to the notification (shared by admins on Reddit and reported by BleepingComputer and Neowin on November 21, 2026), the flaw was introduced on July 23, 2026 and found by Microsoft on November 9. It could have let an unauthorized party read tenant and service-monitoring metadata and potentially modify Connect Health monitoring configurations. Microsoft deployed a server-side fix on November 19, finished its investigation on November 20, and says: "unauthorized read-only access to tenant-level metadata associated with your organization had occurred. We did not identify any unauthorized configuration changes." Exposed for notified tenants: email addresses of tenant technical contacts and administrator accounts; tenant IDs, display names, initial domain names, country codes and region; Entra license tier and directory synchronization status (including last sync time); Connect Health onboarding status; and an expired certificate (the pksCertificate field) once used to register Connect Health agents with an internal Microsoft service decommissioned in late 2024. NOT exposed, per Microsoft: user passwords, password hashes, authentication tokens, valid cryptographic keys, on-premises server and Active Directory topology details, or Azure subscription information. A Microsoft spokesperson confirmed the fix and notifications to BleepingComputer. UNKNOWN: who exploited it, how many tenants were affected (more than a dozen admins, including MSPs with several customer tenants, reported notices), and when exploitation started. No CVE has been published for this server-side flaw.
Who's affected
Organizations that use Microsoft Entra Connect Health and received a direct advisory in Azure Service Health, plus MSPs managing those tenants. The real risk is follow-on targeting: attackers now know admin and technical-contact email addresses, tenant names and domains, license tier and whether directory sync is enabled, which makes convincing Microsoft-themed phishing, MFA-fatigue and help-desk social engineering against privileged users easier.
✅ Fix / patch
- No customer patch is required; Microsoft fixed the flaw server-side on November 19, 2026. Check Azure Service Health (Health advisories / Security advisories) in each tenant you manage to see whether you were notified.
- Warn admins, technical contacts and the help desk that their names and email addresses may be in attackers' hands. Expect Microsoft-, Entra- or 'tenant security'-themed phishing and calls; verify through known channels.
- Make sure every privileged account uses phishing-resistant MFA (FIDO2 security keys, passkeys or Windows Hello for Business), as Microsoft recommends, and review Conditional Access for admin roles.
- Review your Connect Health configuration and Microsoft Entra audit and sign-in logs for anything unexpected since July 23, 2026. Microsoft's read-only PowerShell reference scripts (aka.ms/ConnectHealth/scripts) show the types of metadata the API returns for your tenant, but only current values, not what was accessed.
- Consider whether exposed technical-contact addresses are shared or personal mailboxes and tighten external-sender protections for them.
🔎 Quick hunt check
Microsoft has published no attacker IOCs, IPs or user agents, and it says no configuration changes were found. These queries look for the follow-on risk and for any unexpected Connect Health changes: (1) Entra audit events that touch Connect Health since July 23, 2026; (2) inbound external email to privileged users with Microsoft/Entra/tenant/security lure wording since November 19, with click data in Sentinel.
MITRE ATT&CK: T1589.002, T1590, T1566, T1078.004
IOCs
No IOCs (IPs, domains or hashes) published by the sources yet.
No public IOCs. Microsoft has not shared attacker IPs, user agents or identities. T1589.002 and T1590 describe the reported exposure of email addresses and tenant information; T1566 and T1078.004 reflect the expected follow-on phishing/account-takeover risk covered by the queries, not observed attacker activity.
Hunt queries
Starting points built only from published indicators. Field names vary; tune before use.
```Microsoft Entra Connect Health API breach: Connect Health audit changes since Jul 23 + Microsoft/Entra-themed lures to admins since Nov 19 (exposure check, no attacker IOCs published)```
index=* earliest=07/23/2026:00:00:00 (sourcetype="azure:aad:audit" OR sourcetype="ms:aad:audit" OR sourcetype="o365:management:activity" OR sourcetype=*mail* OR tag=email)
| eval op=coalesce(operationName, activityDisplayName, Operation, ""), svc=coalesce(loggedByService, Workload, ""), subj=coalesce(subject, Subject, ""), sender=lower(coalesce(src_user, sender, from, SenderAddress, "")), rcpt=lower(coalesce(recipient, RecipientEmailAddress, ""))
| eval hit=case(match(op." ".svc, "(?i)connect\s*health|ADHybridHealth"), "connect_health_change", _time>=strptime("2026-11-19","%Y-%m-%d") AND match(subj, "(?i)(entra|azure ad|microsoft 365|tenant|connect health|global admin|admin(istrator)? account)") AND match(subj, "(?i)(security|breach|incident|verify|suspend|action required|unusual|password|mfa)") AND NOT match(sender, "@(microsoft|microsoftonline|azure)\.com$"), "admin_lure")
| where isnotnull(hit) ``` scope admin_lure to your privileged/technical-contact recipients with a lookup if available ```
| stats count min(_time) as first_seen max(_time) as last_seen values(op) as operations values(subj) as subjects values(rcpt) as recipients by hit, sender
| convert ctime(first_seen) ctime(last_seen)
| sort hit - count// Microsoft Entra Connect Health API breach: Connect Health audit changes since Jul 23 + Microsoft/Entra-themed lures to privileged users since Nov 19 (exposure check, no attacker IOCs published)
let FlawIntroduced = datetime(2026-07-23);
let FixDeployed = datetime(2026-11-19);
let CHChanges = AuditLogs
| where TimeGenerated >= FlawIntroduced
| where LoggedByService has "Connect Health" or OperationName has "Connect Health" or tostring(TargetResources) has_any ("Connect Health", "ADHybridHealthService")
| project TimeGenerated, Hit = "connect_health_change", Detail = OperationName, Actor = tostring(InitiatedBy.user.userPrincipalName), Result;
let Admins = IdentityInfo
| where TimeGenerated > ago(14d)
| where AssignedRoles has_any ("Global Administrator", "Privileged Role Administrator", "Security Administrator", "Exchange Administrator", "Hybrid Identity Administrator", "User Administrator")
| distinct AccountUPN = tolower(AccountUPN);
let Lures = EmailEvents
| where TimeGenerated >= FixDeployed and EmailDirection == "Inbound"
| where tolower(RecipientEmailAddress) in (Admins)
| where Subject matches regex @"(?i)(entra|azure ad|microsoft 365|tenant|connect health|admin)" and Subject matches regex @"(?i)(security|breach|incident|verify|suspend|action required|unusual|password|mfa)"
| where not(SenderFromDomain in~ ("microsoft.com", "microsoftonline.com", "azure.com"))
| join kind=leftouter (UrlClickEvents | where TimeGenerated >= FixDeployed | project NetworkMessageId, ClickedUrl = Url) on NetworkMessageId
| project TimeGenerated, Hit = "admin_lure", Detail = Subject, Actor = SenderFromAddress, Result = iff(isnotempty(ClickedUrl), strcat("clicked: ", ClickedUrl), "no click");
union CHChanges, Lures
| order by Hit asc, TimeGenerated desc⏱ CISA deadline
Not in CISA KEV as of Nov 22, 2026.
Sources
- Microsoft: Entra Connect Health read-only reference scripts (aka.ms/ConnectHealth/scripts), linked in Microsoft's Nov 2026 tenant notification
- BleepingComputer: Microsoft Entra Connect Health flaw exploited to steal tenant data (Nov 21, 2026; quotes Microsoft's notification and spokesperson)
- Neowin: After outage, Microsoft confirms another July bug that led to Entra data breach (Nov 21, 2026)
Primary source note: Microsoft has not published a public advisory or blog post; it notified affected tenants privately through Azure Service Health and confirmed the fix to BleepingComputer. The only public Microsoft artifact is the read-only Connect Health PowerShell script repository referenced in that notification. Facts come from the notification as shared by admins and reported by BleepingComputer and Neowin.
Related playbooks
- FBI NCIC breach: the national crime database was reportedly hacked. What we know, what we don't, and the FBI's second breach in two months · Breach · 2026-11-22
- OpenRouter AI inference provider breach: Dragos canaries caught months of prompt theft, stolen AWS keys abused; OpenRouter drops 3 providers · Breach · 2026-11-22
- Bell Canada breach: 'agentic' threat actor stole enterprise customer data from a circuit-provisioning Maximo pre-production system · Breach · 2026-11-21
Get the SockAnalyst Weekly Brief free, every Thursday.
One email a week with the KEV deadlines, patches and hunt queries that matter. New playbooks still post here on the site every morning.
Subscribe free
SockAnalyst