Subscribe
0%
Illustration for: Microsoft 365's Exchange outage runs into day two: what broke, what Microsoft says and how to claim SLA credits

Microsoft 365's Exchange outage runs into day two: what broke, what Microsoft says and how to claim SLA credits

An Exchange Online failure that began Thursday at 16:17 UTC knocked out Outlook and a long list of dependent Microsoft 365 services worldwide. Microsoft reports "encouraging recovery progress" but no root cause. Here's what admins should tell users, and how to document the outage for a service credit.

Listen to this article0:00 / 7:57

Microsoft's Exchange Online, the mail service behind Outlook, has been failing for users around the world since Thursday afternoon UTC, and the outage was still not fully fixed on Friday. In a status update at about 12:00 UTC (6 a.m. Central) on November 13, Microsoft reported "encouraging recovery progress", with service in Europe and Japan becoming more stable and North America improving, according to NBC Chicago. It warned that some users may still see intermittent problems and gave no time for full restoration.

Update, November 14: Microsoft has now given a preliminary cause. In a Friday statement it called the outage "a technical issue not related to cyber or AI / agents", introduced during an earlier network capacity upgrade and exposed by a later network change, PCWorld reports. By about 16:57 UTC Friday it reported "broad recovery and stability" and said mail queues were approaching normal, but it had not declared the incident closed and was still restoring the remaining customers and regions, per TechRadar's live coverage. Microsoft also apologised and repeated that a full post-incident report will follow its review. Kentik's Doug Madory put the disruption at about 18 hours before traffic returned, TechCrunch reported.

Update, November 16: The outage had a physical-safety consequence in Germany. At the Hotel Ling Bao at the Phantasialand theme park in Brühl, a pool's automatic chlorine dosing system overdosed the water and chlorine gas escaped. The system's safety monitor tried to warn staff by email, but the alert never arrived because Exchange Online was down, Notebookcheck reports, citing broadcaster WDR. The fire brigade sent about 40 responders. Accounts of the toll differ. Notebookcheck says 15 guests were affected and four went to hospital. Antenne NRW, citing police and fire services via upday, reports 13 injured and five taken to hospital, one with moderate injuries. The two outlets also disagree on the date: Notebookcheck places it on Friday afternoon (November 13), while Antenne NRW says November 14. According to Antenne NRW, the alarm was raised only after guests complained of breathing problems. The pool was drained, refilled and has since reopened. The lesson, which Security Affairs also draws: a safety alarm shouldn't depend on a single cloud email channel. It needs a local alarm, or a second path such as SMS, that works when the cloud doesn't.

Why it matters: because so much of Microsoft 365 leans on Exchange, this was not just an email outage. Teams calendars, Copilot, Defender's Safe Links, Purview, Intune and parts of the admin center were all listed as affected. At roughly 20 hours of disruption so far, many tenants will be past the thresholds in Microsoft's uptime guarantee, but service credits are not paid automatically.

Timeline (UTC)

Pieced together from Microsoft's status posts as compiled by LazyAdmin, ITdaily and ABC:

When What happened
Thu 14:45–15:53 A separate Teams incident (TM1550488) in the Americas; resolved
Thu 16:17 Exchange Online incident EX1550543 begins; mailboxes won't open
Thu 18:26 MO1550821 opened for services that depend on Exchange
Thu ~19:00 Microsoft pauses deployments and config changes, shifts traffic, restarts servers
Fri 00:00 Servers brought back overnight go unresponsive again once traffic returns
Fri ~05:00 Microsoft says some machines added to absorb load are "stuck in a bad state"
Fri ~07:45 Microsoft reports positive stability results after patching part of the infrastructure (ABC Australia)
Fri ~12:00 "Encouraging recovery progress"; Europe and Japan stabilising

At the peak on Thursday, Downdetector logged more than 130,000 Outlook problem reports in the US and over 48,000 for Microsoft 365, Newsweek reported. ITdaily says reports came in from 74 countries and territories. Those are user reports, not a count of affected people.

What Microsoft says went wrong

There is no confirmed root cause yet. Microsoft's explanations have shifted as it investigated:

  • Early on, it said a large number of servers became unexpectedly unavailable, pushing load onto the rest, and that the load was not caused by more traffic.
  • It then said servers were failing automated health checks, and that a process on them that sets up secure connections had become overwhelmed.
  • Overnight it began rolling back a recent network configuration change that it said correlates with the start of the impact, after earlier saying it had found no such change. On Friday it tied the incident to an earlier network capacity upgrade (see the update above); the formal root cause still awaits the post-incident report.
  • Microsoft says it has seen no sign of malicious activity, and Australia's Home Affairs Minister said he had been told it was not a cyber event, per ABC Australia.

Who was hit

Microsoft's impact list (reproduced by WindowsForum) covers 19 services, including Teams scheduling, Copilot Chat, Defender for Office 365, Purview, Intune, Planner, Graph mail and calendar calls, and some Windows 365 Cloud PCs. Teams chat and meetings and SharePoint files largely kept working.

The knock-on effects reached well beyond IT. ITdaily reports that Anthropic's and OpenAI's Microsoft 365 connectors stopped working. In Australia, the government activated a digital resilience agreement with Microsoft and some regional NSW service centres closed, according to ABC Australia. New York City's emergency management office warned of delayed email to city agencies, ABC News reported.

What admins should do now

  1. Tell users not to resend mail. Microsoft has extended the email timeout window to 48 hours so queued messages are held rather than bounced, and says clearing the backlog is its priority.
  2. Track the right incidents. In the admin center, go to Health › Service health and look up EX1550543 and MO1550821. If the admin center won't load, use status.cloud.microsoft or @MSFT365Status on X.
  3. Don't troubleshoot endpoints. This is a server-side failure. WindowsForum relays Microsoft's advice not to sign out of, clear or reinstall the Power Apps mobile app while it's ongoing.
  4. Warn about Safe Links and scams. A blocked Safe Links URL during the outage doesn't mean the link is malicious. Expect phishing and fake "Microsoft support" calls themed on the outage.
  5. Audit anything that alerts only by email. Building, pool, server-room or equipment alarms that send their warnings through Exchange went silent during this outage (see the Phantasialand update above). Give each one a second channel that doesn't depend on Microsoft 365.
  6. Don't call it resolved early. Microsoft has said repeatedly that recovered servers fell over again under load.

Getting a service credit

LazyAdmin summarises the Microsoft 365 SLA terms like this: Exchange Online guarantees 99.9% monthly uptime, with a 25% credit below 99.9%, 50% below 99% (about 7.2 hours of downtime in a 30-day month) and 100% below 95% (about 36 hours). Uptime is measured in user-minutes, so a tenant's actual figure depends on how many users were down and for how long.

Credits aren't automatic. You file a service request in the admin center with the incident IDs, start and end times and the number and location of affected users. LazyAdmin puts the deadline at December 31, 2026; credits go against a future invoice. CSP customers should ask their partner, who may file for them.

What to watch: Microsoft formally closing EX1550543 and MO1550821 (the end times matter for your credit claim), and the post-incident report, which should explain how a capacity upgrade took down servers worldwide.

Sources

  1. NBC Chicago — When is Microsoft Outlook expected to be back up? Latest update as some emails appear up and running nbcchicago.com
  2. LazyAdmin — Microsoft 365 Outage: Current Status and SLA Credits lazyadmin.nl
  3. WindowsForum — Exchange Online Outage EX1550543: 19 Microsoft 365 Services Affected and What Admins Should Track windowsforum.com
  4. ITdaily — Microsoft 365 outage: Outlook and Copilot sometimes appear online itdaily.com
  5. ABC News — Microsoft Outlook outage updates abcnews.com
  6. ABC News (Australia) — Live: Microsoft outage hits users worldwide, affecting Teams, Outlook and Copilot abc.net.au
  7. Newsweek — Is Microsoft Outlook still down? newsweek.com
  8. PCWorld — Microsoft: AI, hackers not to blame for Outlook services outage pcworld.com
  9. TechRadar — Microsoft outage live techradar.com
  10. TechCrunch — Microsoft's latest email outage turned into a full-blown global cloud crisis techcrunch.com
  11. Notebookcheck — Swimmingpool-Chlorüberwachung per E-Mail: Verletzte durch Microsoft-Ausfall notebookcheck.com
  12. Antenne NRW — Chlor-Unfall im Phantasialand: Neue Zahlen sorgen für Verwirrung antenne.nrw
  13. Security Affairs — The Microsoft Outage Left 15 Hotel Guests Poisoned by Chlorine securityaffairs.com

#Microsoft#Microsoft 365#Exchange Online#Outlook#Teams#outage#SLA