Microsoft cloud and Azure companies hosted on the West Coast of the US went down for hours on Thursday when community connectivity failed. Though companies operating solely inside Microsoft’s West US cloud area have been unaffected, any visitors coming into or leaving the amenities was affected.
Microsoft has now revealed a Preliminary Publish Incident Assessment (PIR) of the incident, reporting that connectivity was misplaced for 5 hours between 14.44 UTC (7.44 a.m. Pacific Time) and 19.41 UTC on July 23. The issue was triggered when a set of IP routes was eliminated in error whereas isolating a tool for routine upkeep.
Earlier than beginning the upkeep work, Microsoft checked that no less than one of many two redundant paths to the power remained operational. When it got here to beginning the work, nonetheless, automated techniques included some extra units within the perimeter to be remoted, and eradicating some IP routes that had not been included within the preliminary evaluation.
Clients found the issues in a short time, and engineers recognized the difficulty throughout the first hour and began to reconnect companies. Microsoft stated the disruption had been attributable to some “current fiber upkeep exercise”.
To reduce the danger of disruption from such errors sooner or later, Microsoft suggested organizations dealing with mission-critical knowledge to think about a multi-region strategy.
The Azure outage was the second vital one to hit Microsoft this yr. In February, there was a 10-hour disruption to US West and US East areas.
This text first appeared on Community World.
