The huge, global outage at Microsoft had been lying in wait for months, the company’s official explanation has revealed.
On Thursday, Microsoft Outlook and other products went offline, and were not restored for around a full day. The outage led to global chaos, affecting everything from hospitals and police forces to schools, as well as grounding work across the world to a halt.
Now, the company has revealed the “root cause” of the problem, in its first report into how it was able to happen. It said it was caused by “three factors” that together took down a key piece of Microsoft’s network – and that the seeds of the trouble had actually been planted in July.
Months ago, Microsoft issued an update to the servers that power Microsoft 365. It started rolling out in July, and continued safely over “several weeks” during which there appeared to be no problems, it said.
For more than three months, the change appeared to have rolled out fine. No issues were fine and testing suggested that there were no problems.
But it did in fact include a bug that meant that servers would become overwhelmed and unresponsive when they were dealing with high amounts of load, Microsoft said. the company had not spotted it because the testing had not looked for what would happen with large amounts of traffic, and because it was especially problematic on newer and more powerful hardware, which was not being used during that process.
Then, Microsoft’s network was hit by the kind of high load that would trigger those problems. About 15 hours before the outage began, Microsoft Teams switched from its usual network to the one that also powers Exchange Online, which is used to deliver emails through Outlook.
That in itself would normally be fine. But at the same time, around 14 per cent of Microsoft’s server capacity was out of service because of an unrelated issue, which Microsoft had scheduled to return to service through its routine processes.
Each of those issues on their own would normally be fine. But the two central problems – an increased load and a reduced server capacity to deal with it – meant that Microsoft’s systems ran into the problem that had been lying in wait from July.
That meant that requests built up and the servers stopped responding, on Thursday, during the peak time for traffic in the US. Usually, Microsoft’s systems would automatically switch over to other servers that were working, but that in turn meant that the other servers were more likely to run into that same issue.
That led to a cascading series of problems, as more traffic was sent to fewer servers, meaning that those servers too were likely to go offline. In less than two hours, that problem had broken servers right across Microsoft’s network.
At its worst, 80 per cent of Microsoft’s servers were unresponsive, it said. The system was unable to cope with the traffic and so broke entirely.
That meant that a whole host of Microsoft services – Outlook, but also other products such as Teams, OneDrive, SharePoint and Office apps – failed. Some of those apps switched from the broken network and so could be restored sooner, but Outlook could not do that.
Microsoft’s attempts to fix the problems were hit by issues of their own. As soon as servers switched back online, they were hit by a huge amount of traffic that overwhelmed them and meant that they switched back off again.
What’s more, because the problem had actually resulted from a bug that had been in place for several months, Microsoft was not able to easily roll it back again. That meant that the fix was delayed further, and meant that Microsoft had to build an update to fix the issue – which meant that it had to take groups of servers out of service, update them, and restart them.
Charles Lamanna, Microsoft’s executive vice president of Copilot, agents, and platform, said that the company was “deeply sorry for the impact to all our customers and recognize the impact that this has on their normal business and operations”. Microsoft also said that it would be releasing a more detailed, final report on what happened later this week.
Read MoreMicrosoft confirms services fully restored after major global outage
Outlook issue causes chaos around the world
Microsoft ‘deeply sorry’ for disruption caused by major global outage
Trump’s old tweets redirect to Epstein files after hijacking
Shock report finds AI-generated child sex abuse images soar 8,400% in two years



