A major Amazon outage on March 5, 2026 generated over 20,000 Downdetector reports peaking at 20,804, with 50 percent tied to checkout failures; Amazon attributed the incident to a software code deployment and resolved it by 9:05 p.m. ET (Ars Technica, March 6 2026 update). The original coverage documented user impacts and the Amazon support acknowledgment on X but did not address recurring deployment errors across AWS history.

Primary source synthesis with the AWS postmortem for the December 7 2021 outage that disrupted Netflix, Disney Plus and Roku services after a network configuration change in US-EAST-1 (AWS Premium Support, 2021) and the official summary of the February 28 2017 S3 outage caused by an erroneous command removing capacity (AWS Message 41926, 2017) reveals a documented pattern of code and configuration failures. Initial Ars Technica reporting omitted these direct parallels and underreported the concurrent smaller AWS complaints that indicated broader service ripple.

Synergy Research Group data from 2025 shows AWS holds 31 percent of global cloud infrastructure spend; Gartner 2024 tracking logged over 60 percent of enterprises experiencing multiple annual cloud incidents. These elements, combined with the current event, demonstrate how centralized cloud providers create single points of failure that propagate across e-commerce, streaming and enterprise operations when routine deployments lack sufficient safeguards.