For two days this week, a lot of inboxes simply would not open. A widespread Microsoft 365 and Exchange Online disruption began on Thursday, November 12, 2026, and was still being cleaned up on Friday, November 13. Microsoft tracked the mailbox incident as EX1550543 and the broader Microsoft 365 incident as MO1550821. Affected connection methods included Outlook on the web, MAPI, Exchange ActiveSync and REST APIs, so the mailbox service itself was hard to reach.
If you run hosting servers, your websites may have been fine. That is the problem worth sitting with. Web hosting and email often get sold as one "cloud," but they fail on different clocks. When mail depends on a single provider, a bad afternoon there is a silent mailbox, even while the company website still loads.
What happened
Microsoft said a number of servers became "unexpectedly unavailable," which put extra load on the systems still running. News reports, including the BBC, also described failed automated health checks. The company said there was no evidence of malicious activity. It has not been presented as a cyberattack or a data breach.
Microsoft has not published a count of affected mailboxes. The BBC reported that Downdetector logged about 140,000 Outlook problem reports from 74 countries on Thursday evening, likely a large undercount because the site only counts people who choose to report. Microsoft's figure of more than 430 million Microsoft 365 app users describes the size of the product, not how many people lost service.
By late morning Friday, November 13, 2026, Eastern Time, Microsoft said it was seeing "sustained improvements in service availability" and that email queues were "approaching normal levels," while warning that some users may still have intermittent access issues or delayed delivery. Its public status page still listed a service degradation when we checked it. Microsoft had also extended the email timeout window to 48 hours so delayed messages would keep retrying, and it has promised a post-incident report after mitigation is complete.
The single-provider problem sits in your MX
Mail does not "find" Outlook on its own. It follows the MX records for the domain. Those records name the hosts allowed to accept mail, and they have priorities. If every MX target, and every way your users connect to mail, lives inside one vendor's network, incoming mail has nowhere else to go.
A backup MX can help, but only with clear eyes. A lower-priority MX on a genuinely independent mail system can accept messages when the primary host is unreachable and hold or forward them later. It does not help if the backup is just another name in the same cloud, and it may not fail over cleanly if the primary is still reachable but hanging or answering slowly. A backup MX is also not a mailbox your staff can log into while the primary is dark. For day-to-day continuity, an independent secondary mailbox, with contacts exported and a phone or web-form path, is usually the more useful tool.
Do not panic-edit DNS in the middle of an outage. MX changes are delayed by TTL, mail already in flight can split between paths, and if the original provider is offline you may not be able to pull existing mail out of it until it recovers. Thexyz's write-up of this week's outage is worth reading before anyone starts rewriting records under pressure: Microsoft 365 Is Down Again. FSO News also covered the second-day disruption and the cost of a single provider: Microsoft 365 Outage Stretches Into a Second Day. For a practical walkthrough of a backup mailbox, see the Webmail Wiki article on keeping business email running.
What to check once mail is moving again
- List the MX records for every domain you host or manage. Note the targets, priorities and TTLs. If they all point at one cloud, write that down as a dependency, not as a given.
- Check SPF, DKIM and DMARC. Any secondary sending path you might use later has to be represented there, or mail you send from it will look forged.
- Confirm that a second mailbox actually works at a provider that does not share Microsoft's infrastructure, and that someone on the team can sign in today. A leftover account nobody can find is not a backup.
- Export a current copy of contacts and keep a phone or ticket path that does not depend on Outlook.
- Review your monitoring. If the only status page you watch is the vendor's, you will learn about a global failure the same way your customers do.
- Be honest about failure domains if you offer mail on the same servers as websites. Shared hosting mail and a Microsoft 365 tenant fail differently. The useful question is whether customers have more than one working path.
Thexyz, which sells Canadian-hosted independent email and has a commercial interest in this conversation, argued this week that professional mail does not have to live inside a full Microsoft 365 suite. That is a product argument, disclosed as such in their article. The hosting lesson does not require a provider change today. It does require knowing, in writing, where your MX points, how long a change would take, and what you would use if that path went dark again. No mail host is immune to downtime. The goal is a design that still lets you send and receive something when one vendor is having a bad day.
Sources
BBC News live, Global Microsoft outage: https://www.bbc.com/news/live/cq5y10ge3dp3t
Microsoft service health status: https://status.cloud.microsoft/m365
Thexyz, Microsoft 365 is down again: https://www.thexyz.com/blog/microsoft-365-is-down-again-how-many-outages-price-hikes-and-privacy-trade-offs-are-too-many/
FSO News, Microsoft 365 outage stretches into a second day: https://www.fsonews.com/2026/11/microsoft-365-outage.html
Webmail Wiki, Outlook down? Set up a backup mailbox: https://www.webmail.wiki/2026/11/outlook-down-backup-email.html
Post a Comment