During AWS outage, one engineer kept enterprise services up with just 22 min downtime

generativist · x · 2026-09-12

generativist recounts keeping enterprise services running with Fable through the AWS mega-outage: migrating across zones ("yolo fail"-style, switching ports and locations), keeping redundant instances up and routing around shutdowns, for only 22 minutes of total downtime all day.

A follow-up notes service recovered right around the time of loudly complaining on Twitter, suggesting an automated AWS system had gotten into a really bad state. A rare public post-mortem-style account of riding out the outage.

Original post →

More from Infra

Infra channel →