General Tech

Three small failures combined to take down Microsoft 365

A Microsoft 365 outage was caused by a hidden bug in the company's cloud gateway that was introduced in July, combined with shifting traffic and reduced server capacity. The bug caused CPU and memory resources to become exhausted and unresponsive, leading to a cascade of failures that affected Microsoft 365 and Exchange infrastructure for 24 hours. The incident is attributed to a change deployed in July that created the defect, which was not discovered in pre-release testing due to the use of less powerful hardware. The bug was exposed when traffic shifted during standard network maintenance and upgrades, and the reduced server capacity meant that the system was unable to handle the increased load. Microsoft has released a preliminary post-incident report outlining the causes of the outage, which is available to users with Microsoft 365 admin credentials. The incident is covered under incident IDs EX1550543 and MO1550821, and the University of Pennsylvania has archived Microsoft's service-health updates.

Read the full article at arstechnica.com →