News

Microsoft Reveals How Over 80% of Servers Failed During Global Outage

Microsoft revealed that over 80% of its networking gateway servers became unresponsive during a global outage. A software defect, increased traffic, and reduced server capacity triggered the cascading failure.

Written By : Somatirtha
Reviewed By : Manisha Sharma

Over 80% of servers behind Microsoft's networking gateway became unresponsive at the peak of last week's global outage that affected Outlook and other Microsoft 365 services, the company said on Monday. Microsoft blamed a software defect introduced in July that remained undetected for more than three months.

Charles Lamanna, Microsoft's Executive Vice President of Copilot, Agents and Platform, said three factors triggered the failure.

A Hidden Defect and Rising Traffic

The first factor was a latent defect that caused servers to become overwhelmed under certain load conditions. It stemmed from a fix prepared specifically for Microsoft 365 networking gateway servers.

The change began rolling out in July and remained in place for more than three months without impact. Microsoft's load testing missed the problem because it focused on weaker hardware; the defect is more severe on newer machines with more processing cores.

About 15 hours before the outage, Microsoft Teams shifted from an alternative gateway to the primary gateway, which also supports Exchange Online. This increased the load. Nearly 14% of the gateway's server capacity was also out of service because of an unrelated issue from the previous day.

Failure Spread Across the Network

The increased load and reduced capacity exposed the flaw. Requests built up during peak traffic hours in the United States, leaving servers unresponsive.

Microsoft's traffic management systems redirected users to healthy endpoints. The redirected traffic put further pressure on remaining servers and triggered a cascading failure. In less than two hours, more than 80% of servers were unresponsive.

Also Read: Microsoft’s Next Xbox Wants to End the Console vs PC Gaming Debate

Recovery Took Longer than Expected

Teams, OneDrive, SharePoint and Office apps moved away from the faulty gateway and were restored earlier. Exchange Online could not do so.

Recovery was delayed because returning servers were overwhelmed by waiting traffic. Microsoft's automated processes were not designed to handle a failure on this scale.

Engineers developed an update to restore the code to its previous state. They took groups of servers offline, applied the fix, and restarted them before returning them to service. Lamanna said they would provide a final post-incident review later this week.

Join our WhatsApp Channel to get the latest news, exclusives and videos on WhatsApp

Bitcoin Miner Jiang Zhuoer Exits DeFi Tokens After BTC Slides Toward USD 62,000

Strategy and Strive Signal Bitcoin Buying Interest as MARA Sells USD 1.1 Billion

Solana Co-Founder Responds to Ethereum Critics After Firedancer Node Runs on Tablet

GuruFun Launches Community Pad on Solana, Putting Community Before the Market

CLARITY Act Faces Uncertain Future as Congress Prioritizes SEC and CFTC Appointments