6.2.4. Service Health, Message Center, and Alerts
💡 First Principle: When something breaks, the first question is whose problem it is: Microsoft's service, an expired credential in one of your connectors, or your own configuration. Tenant status answers the first two. For your own estate, alerting means deciding which signal should reach which person, then wiring that signal to a notification channel. Intune gives you the signals; you choose the channel.
Intune tenant status. Tenant administration > Tenant status has three tabs. Tenant details shows the tenant name and location, the MDM authority, the service release number and licensed users. Connector status covers sync connectors such as Apple VPP and Windows Autopilot and credential connectors such as APNs, sorted with unhealthy first:
| Status | Meaning |
|---|---|
| Unhealthy | The credential has expired, or the last sync was three or more days ago |
| Warning | The credential expires within seven days, or the last sync was more than one day ago |
| Healthy | No expiry within seven days, and the last sync was less than one day ago |
An APNs certificate that expires in five days shows Warning; if nobody renews it, it turns Unhealthy the moment it expires. A Healthy status doesn't guarantee a connector works, though; it only means the credential and sync timing look right.
Service health and Message center. The third tab, Service health and message center, shows only incidents and advisories that affect your tenant, plus the 10 most recent messages. Links open the Microsoft 365 admin center for older items. There, Health > Service health distinguishes an advisory (some users affected, service still available, often with a workaround) from an incident (the service or a major function is unavailable). Planned maintenance doesn't appear in Service health; track it in Message center by filtering on "Plan for change". Email notifications for both are configured only in the Microsoft 365 admin center, under Health > Service health > Customize > Email, with up to two addresses per admin, filtered by incidents or advisories and by service.
Building alerts for your own estate. The outline asks for alerts on compliance drift, enrollment failures and configuration conflicts. Here's how you build each with the tools Intune documents:
| Requirement | How you build it |
|---|---|
| Tell users their device has fallen out of compliance | Actions for noncompliance in the compliance policy: send email (from a notification template), push notification on Android and iOS/iPadOS, remote lock on mobile devices, add to retire list |
| Tell admins about enrollment failures | Diagnostic setting sending OperationalLogs to Log Analytics, then an Azure Monitor alert rule on the operational log table; the Enrollment failures report for investigation |
| Tell admins about compliance drift | Diagnostic setting sending DeviceComplianceOrg to Log Analytics with an alert rule, remembering this data arrives every 24 hours; the Device compliance trends report for review |
| Spot configuration conflicts | Configuration policy assignment failures and Profile configuration status reports, or an alert on the audit and operational data in Log Analytics |
| Alert on Cloud PC provisioning or connection failures | Tenant administration > Cloud PC Alerts > Alert rules: conditions, thresholds, severity (Informational, Warning, Critical), portal or email notification |
The notification details are testable. The default Mark device noncompliant action runs at 0 days and can't be removed. Each other action has a schedule from 0 to 365 days in steps of 0.25 days, so 0.25 means six hours and 1.5 means 36 hours. Emails go to the user's profile email address rather than the UPN and are expected within six hours. Templates live at Endpoint security > Device compliance > Notifications > Create notification. Remote lock isn't supported for Windows, but adding to the retire list is; an admin still retires the devices on that list. Cloud PC alert rules need Windows 365 Enterprise and the Intune Administrator or Windows 365 Administrator role, and they send one email per trigger.
Latency shapes alert design. Operational and audit logs flow to Log Analytics immediately, so enrollment-failure alerts can be close to real time. Compliance and device data arrives once every 24 hours and can take up to 48, so a compliance-drift alert is a daily signal, not a live one. The pipeline may also duplicate data, so alert queries should handle duplicates.
⚠️ Exam Trap: "Email the user when their device becomes noncompliant" and "alert the admin team when enrollment failures spike" need different tools. The first is a noncompliance action with a notification template, inside the compliance policy. The second needs Intune data in Log Analytics with an Azure Monitor alert rule, and that needs an Azure subscription. Cloud PC alert rules apply to Windows 365 Cloud PCs, not to physical devices.
Reflection Question: New iPhones suddenly fail to enroll across the company, and the Microsoft 365 Service health page shows no Intune incident. Where would you look next, and which status values would confirm your suspicion?