Microsoft Struggles to Restore Outlook and Teams After Nine-Hour Outage Hits Thousands

By early Friday morning GMT, Microsoft had been trying for more than nine hours to bring its Microsoft 365 services back, and the fixes it had deployed kept failing. The disruption began Thursday and reached thousands of users of Outlook, Teams, Word, Excel and PowerPoint, along with enterprise tools such as SharePoint and OneDrive. Microsoft's Copilot AI assistant was also affected.

The scale showed up on Down Detector, the outage-tracking site. It recorded a spike in Microsoft 365 problems around 1pm ET, with nearly 130,000 reports of issues from 74 countries.

Microsoft said early on that it had no indication of a cyber-attack. Charles Lamanna, the company's executive vice-president of Copilot, repeated that on X. He wrote that the company regretted the disruption to customers, was working to address it, and would share more as soon as it had it. He confirmed there was no indication the outage was the result of malicious or cyber activity.

The company's working explanation points to infrastructure. Microsoft suggested the problem may lie in the settings applied to the servers that keep Microsoft 365 running. It said part of its infrastructure was experiencing elevated load, which was likely contributing to connectivity errors, and that a number of servers were unexpectedly unavailable because they failed automated checks. Its engineers tried rolling back updates and limiting the traffic sent to affected servers. Each time, the servers became unavailable again once traffic returned. As of 3.30am GMT Friday, Exchange, the back end to Outlook email, was the most affected, with knock-on effects for products that depend on it. Teams chats and meetings were less affected. Microsoft had earlier said it restored Teams problems in North America, though some users still reported trouble.

The company said the servers it had restored were holding stable. It was shifting traffic to alternate routes where possible and streamlining the process of bringing the remaining servers back. It was also assessing whether reverting a small number of recent configuration updates it had identified could speed recovery. Users, it said, may see a gradual recovery as their clients reconnect.

The effects spread beyond Microsoft's own products. OpenAI said it was investigating a problem with Microsoft connectors, in which new connections might be broken and tools might not work as expected. Anthropic said an issue with an upstream provider was affecting its Microsoft 365 connector. The outage also follows another last month, when a failure of Microsoft's Azure cloud platform hit its productivity software and a range of industries worldwide.

Mark Gregory, an associate professor in the School of Engineering at RMIT University, said the outage appears to affect the network layer of Microsoft's systems. He compared it to the configuration problem that brought down Optus's network in 2023. In his account, a configuration fault on some servers forces customers to share a smaller pool. The remaining servers become overloaded and go down, producing a cascading failure that takes time to trace to a network-layer misconfiguration. His assessment is an outside reading of the symptoms, not a Microsoft finding.

Gregory said companies should invest more in testing environments for configuration changes before pushing them to the live network. He said businesses reliant on cloud services should keep alternatives. Many people, he noted, keep a Gmail account to use when Microsoft goes down, and the reverse. Somewhere in the world, he said, someone has been badly hit, with lost productivity. What to watch now is whether the traffic-limiting and rollback efforts finally hold, and what Microsoft says about the root cause once services stabilise.