4.1.5. Network Connectivity and Service Health
💡 First Principle: When users say "Microsoft 365 is slow," the cause is either Microsoft's service or the path between the user and it. Service health tells you about the first; network connectivity insights tell you about the second — so check both before escalating.
Network connectivity — Microsoft 365 admin center > Health > Network connectivity — scores your network (0–100) for the tenant and per office location, and shows insights such as distant internet egress, WAN backhaul and suboptimal DNS. Service Support Administrators configure it; Reports Readers can view it.
Office location data comes from three sources:
- Windows Location Services (off by default) — automatic discovery at city level from Windows devices running a current OneDrive.
- Locations you add with their LAN subnets (egress IPs optional, CSV import supported).
- The Microsoft 365 connectivity test tool, whose results show in the admin center a few minutes after a signed-in test.
The insights rest on Microsoft's network connectivity principles: identify and differentiate Microsoft 365 traffic; egress connections locally (local DNS and local internet egress together); avoid network hairpins; and bypass proxies and traffic inspection for trusted Microsoft 365 traffic. Microsoft publishes its endpoints in Optimize, Allow and Default categories, with changes announced 30 days ahead. The test tool flags egress, proxies or DNS resolvers more than about 500 miles (800 km) away, and checks Teams media quality and Copilot connectivity.
Service health — Health > Service health — shows Microsoft's view of each service. Service Support Administrators and Helpdesk Administrators can view it without broader rights.
- An incident is a critical issue: the service or a major function is unavailable. An advisory is a problem affecting some users with limited impact, often with a workaround.
- Status values run from Investigating through Service degradation or interruption, Restoring service and Extended recovery to Service restored (or False positive), and a post-incident report may follow.
- History covers the last 7 or 30 days. Report an issue lets you tell Microsoft about a problem, and Issues for your organization to act on lists problems detected in your own environment.
- Email notifications (Customize > Email) are set per admin: up to two addresses, incidents and/or advisories, and the services to cover. You can also follow a single issue.
- Planned maintenance and upcoming changes appear in Message center (Health > Message center), not Service health. Message center offers a weekly digest, major-update emails (posted at least 30 days ahead when action is required) and data-privacy messages. The Message center reader role gives read-only access.
⚠️ Exam Trap: If an admin wants to be emailed when Exchange has an advisory, they configure their own Service health email notifications — each admin sets their own preferences. And "find next month's planned maintenance in Service health" is wrong; it's in Message center.
Reflection Question: Users in your Singapore office report slow Teams calls while other offices are fine, and Service health shows no Teams issues. Which admin center page would you open next, and which connectivity principle is most likely being broken?