Tech Meltdown Tracker

Tech's worst days, sourced and scored

Nov 18, 2026·52 incidents in 2026

DownResolved 1

Microsoft's maintenance tooling removes too many network routes and cuts off West US Azure for five hours

A bug in a request-conversion system marked extra devices for isolation during routine maintenance, severing traffic into and out of the region.

Microsoft · Bad update or config change

Post on X

From about 14:44 UTC on July 23, traffic entering or leaving Microsoft's West US Azure region failed, disrupting Azure and other Microsoft cloud services for roughly five hours. Workloads running entirely inside the region largely kept working, according to coverage of Microsoft's preliminary post-incident review.1,2

Microsoft said engineers were isolating a device for routine maintenance when a bug in the request conversion system marked additional devices as part of the event, removing IP routes between the datacenter and the wide-area network. Microsoft had checked beforehand that a redundant path remained, but the automation widened the scope.1,2

Engineers spotted the problem within the first hour and reversed the change. Microsoft said it plans to improve validation and change-management controls and advised mission-critical customers to consider multi-region designs.1,2

“a bug in the request conversion system incorrectly marked additional devices as a part of the maintenance event and caused a set of IP routes to be removed from more devices than intended”
Microsoft, preliminary post-incident review, as quoted by Petri1

Sources

  1. 1
  2. 2
    Microsoft explains why its West US Azure and cloud services failed

    Network World · News report · Jul 24, 2026