More

    The Microsoft 365 Outage: What Microsoft Has Confirmed and What It Has Not

    Microsoft spent most of Friday, November 13, 2026 bringing Microsoft 365 back online after an outage that began the previous evening and left Outlook, Teams, SharePoint, OneDrive and Copilot degraded for about 18 hours. The company has apologised, ruled out a cyber-attack and promised a fuller account. Several practical questions are still open.

    How the incident unfolded

    Monitoring firm ThousandEyes recorded availability problems across Microsoft 365 services from roughly 17:30 UTC on November 12, with connection resets, timeouts and HTTP server errors, plus significant traffic loss inside Microsoft’s network. The outage tracker Downdetector logged close to 130,000 reports across 74 countries, covering Word, Excel, PowerPoint, Outlook and Teams as well as enterprise services such as SharePoint and OneDrive.

    Recovery was gradual rather than a single flip back to normal. Microsoft said it took affected systems offline for fixes and returned them in stages, and that services were recovering shortly before midday GMT on Friday. As of mid-afternoon in Toronto on November 13, the public status page still described a service degradation for Microsoft 365, with users potentially unable to connect to multiple services.

    What Microsoft has said about the cause

    Microsoft’s public position has three parts, and they are worth separating.

    • It was not an attack. Microsoft said the outage was a technical issue and that it had found no indication of malicious activity or a link to AI agents.
    • It followed infrastructure work. In statements reported by CNBC, Microsoft said the connectivity disruption stemmed from an issue during an earlier network capacity upgrade, and that the underlying cause had been addressed.
    • The health-check layer was part of it. Microsoft described problems with a component responsible for processing server health, and said it validated an alternative service health check method at around 4 p.m. ET on Thursday before applying that change more broadly.

    The Guardian reported that Microsoft suggested the outage may have been caused by settings applied to the servers responsible for keeping Microsoft 365 running. Those accounts are consistent but not identical, which is why a final root-cause report matters.

    What stayed up, and what did not

    Azure was not listed as degraded on Microsoft’s status page during the incident, which is a useful distinction: the outage hit Microsoft 365 workloads and identity-adjacent services rather than the whole cloud. Within 365, the failure was patchy. Some organisations lost mail flow entirely; others saw intermittent sign-ins and slow file access. Government bodies including the US House of Representatives and French municipal authorities reported problems, and healthcare and police services in the United Kingdom were affected.

    Third parties that lean on Microsoft connectors were dragged in. OpenAI said it was investigating broken Microsoft connections, and Anthropic said an upstream provider was affecting its Microsoft 365 connector.

    What it means if your mail was delayed

    Microsoft extended the retry window for delayed messages to 48 hours and said email queues were approaching normal levels as recovery progressed. In practice that means a message sent during the outage can still arrive a day or more later. Resending it repeatedly creates duplicates; waiting, then checking the Sent folder and the recipient’s spam quarantine, is the cheaper path.

    The questions that are still open

    Three things are not yet established. Microsoft has not published the full incident report it promised, so the exact sequence from capacity upgrade to global degradation is still its own account rather than a documented timeline. It has not said whether service credits will be offered automatically to affected customers. And it has not explained how a fault in server health checking spread to so many workloads at once, which is the part that matters for anyone reviewing their own redundancy plans.

    What to do differently next time

    The practical lesson is unglamorous: know where your status pages are before you need them, keep an out-of-band way to reach your team, and avoid password resets during an authentication wobble. Our step-by-step guide on how to tell whether Microsoft 365 is down or the problem is on your end covers the check order in detail.

    A plaza and walkways between buildings on the Microsoft campus in Redmond
    A plaza on the Microsoft Redmond campus. Photo: Jonathan Schilling, Wikimedia Commons (CC BY-SA 4.0).

    Sources

    Latest articles

    spot_imgspot_img

    Related articles

    Leave a reply

    Please enter your comment!
    Please enter your name here

    spot_imgspot_img