Microsoft 365 / Service health and Message center: check outages and changes
If Microsoft 365 appears to be having an outage, first open Health → Service health in the admin center. Service health shows current incidents and advisories; Health → Message center helps administrators prepare for feature changes and required actions. Use them together to separate current incident response from preparation for future changes.
Differences between Service health and Message center
| Item | Service health | Message center |
|---|---|---|
| Main Purpose | Check current incidents, performance issues, and recovery progress | Check for changes in functionality, end of support, administrative tasks, and planned updates |
| Administrator's questions | Is the problem occurring now a Microsoft service outage? | Are there any changes our organization should prepare for in the future? |
| Practical use | Share the scope of impact, start time, status updates, and workarounds | Schedule of implementation, affected users/services, and assignment of necessary tasks |
Microsoft’s Service health guide and Message center guide describe the supported controls. The notices shown depend on your organization, service eligibility and role.
Order of checks when an outage is suspected
- Open Health → Service health in the admin center and compare the affected service and region to user symptoms.
- Open related incidents or advisories to check start time, scope of impact, recent updates, and whether a workaround exists.
- Check the impact within your organization separately. The absence of incidents in the service status does not mean there are no problems with organization settings, network, or specific users.
- If you need to contact support, record the symptoms, start time, affected user range, and what has already been checked. If there is an official incident ID, include it as well.
- After a recovery notice is posted, also recheck actual user behavior. A Service restored status does not guarantee immediate normal operation for all users.
Compare the incident notice with your organization’s symptoms
An incident is a critical issue that makes a service or a major function unavailable. An advisory concerns a known issue affecting some users while the service remains available; it can be intermittent or limited in scope and may have a workaround. A notice with the same service name does not prove that a user’s reported problem belongs to that incident. Compare the user impact, start time, affected function or region and the symptoms in your organization.
Investigating, Service degradation, Restoring service and Service restored describe different stages. Read the latest update time and action guidance rather than copying only the first notice. If Microsoft has not supplied an estimated recovery time, do not invent a completion time for internal updates.
For example, if a Teams meeting connection issue is reported, check whether it occurs on other users, devices, or networks, and compare it with the scope of the announcements. Even if the announcement matches, you must check whether there were any simultaneous changes in organizational policy. By leaving the incident number, internal impact scope, last check time, and the next person responsible for verification, you can reduce duplicate investigations by multiple support staff.
Service health alerts and issues without a notice
- Open Service health → Customize → Email, enable Send me email notifications about service health and choose incidents, advisories and the services to monitor. Verify the recipient addresses and save. Microsoft allows up to two email addresses per administrator account; confirm the successor’s own preferences when handing over this responsibility.
- If only updates for a specific incident are needed, use individual notification paths like Manage notifications for this issue in the incident details.
- If symptoms exist but no matching notice appears, record the service, timestamps with time zone, affected users and errors, then use Report an issue. Check its status under Reported issues. Submitting a report does not guarantee that Microsoft will identify a service-side incident; continue the internal investigation as needed.
- If network, account, or permission issues are suspected, conduct an internal investigation within the organization simultaneously. For emails, use message trace, and for logins, check separately using sign-in logs.
Notification reception settings may vary by the account of the person in charge, so check the new person's reception status during handover. Also, verify that notices are not hidden due to custom filters in the service you are viewing. Do not assume that an incident did not occur just because there was no email notification.
How is the message center operated?
In the message center, you can check notices about new features, changes in default behavior, and actions required by administrators. First, confirm whether the service actually applies to your organization, and write down the implementation schedule, user impact, and whether documentation, training, or policy changes are needed. For important notices, designate a responsible person and a completion date for tracking. Make use of filters and notification settings, but it is better not to rely solely on notifications and skip regular checks.
How to turn change notices into actual actions
In Message center, filter the services your organization uses and review tags such as Plan for change, Major update, User impact and Admin impact. Check Act by for notices that require action by a deadline. Prioritize the actual required work and due date; announcements do not necessarily apply to every organization or arrive everywhere on the same day.
| Items to record | Questions to confirm |
|---|---|
| Message ID and update time | Is it the same notice that has changed, or is it a new notice? |
| Target · Distribution Schedule | Which services, platforms, or users will be affected and when? |
| Action · Person in Charge | Who changes what in the app, settings, or user guidance? |
| Internal Completion Deadline | Is there enough time to finish testing and approval before the official deadline? |
| Confirmation Result | Do the actual operations and the documentation match after the change? |
Some supported feature announcements show Status for your org: Scheduled means the release is planned but not yet available to users in the organization; Rolling out means it is reaching some applicable users; Launched means it is generally available to the applicable users. The field is available only for limited announcements. Its absence does not prove that a feature has not been deployed, and phased rollout can produce different screens for users in the same organization.
Marking a notice as read or archiving it in your list does not record completion of the organization’s actual change work. Track important notices in an internal task record with an owner, due date and verification result. When the schedule or guidance changes, reopen the latest content under the same message ID.
At the top of Message center, open Preferences. Use Custom View to choose the services displayed and Email to configure the weekly digest and major-update notifications, then save. Service health incident alerts and Message center change emails are separate preferences. Microsoft sends Message center digest emails and emailed posts in English; the portal’s preferred-language display does not translate those emails.
When the menu is not visible or there are no notices
If a menu is missing, check the account’s role and expand the admin center navigation. Roles such as Service Support Administrator can view Service health. If you cannot access the admin center, check the public Microsoft service status page for known issues affecting access. This public page does not list every problem specific to your organization.
For staff who only need to read and share Message center posts, consider the Message center reader role. Data privacy notices are limited to the Global Administrator and Message center Privacy reader roles; do not grant broad administration solely to read a notice when the limited role fits. Some other admin roles, such as Reports reader, do not provide Message center access. After recovery, repeat the original failed task in the relevant web or desktop app. If a temporary workaround was used, verify the approved normal path when it is restored and record the result.




