Most of us sense we’re in an affordability crisis these days. If you’re like me, you’re helpless and complacent at the checkstand even when it feels like you’re being mugged. But being billed for billions—or even trillions—more than you owe on web hosting would snap anyone out of their affordability daze.
Amazon Web Services users around the world have noticed one such glitch:
I just saw $1.5 trillion on my AWS bill and my soul left my body
— Bharath
Bharath, an X user based in India, showed off what looks like a $1,499,659,180,107 cost statement and writes, “my soul left my body.” That statement says Bharath’s total is up by 744,728,201,771% this month, which means, by my math, the previous month’s bill was about $200.
According to the Guardian, a marketer named Dan Harvey, working for an educational nonprofit in the U.K. said he “almost had a heart attack” after seeing a bill climb from 43 cents last month to $7.8 billion this month—and the month wasn’t even over. Harvey added to the Guardian that he had to get on the phone with tech support and “have a real dig around,” to get to the bottom of things. Amazon did not apparently return the Guardian’s request for comment.
This has been resolved, according to Amazon, which writes that on July 16 and 17, “customers received erroneous budget and cost anomaly detection alerts, and saw inflated estimated cost and usage data in the Billing and Cost Management Console and the Cost and Usage Reports.” The amounts are “inaccurate” and “did not affect customer invoices,” Amazon writes, but everything has apparently been restored to normal.
An update Saturday on the AWS service health dashboard lays out what happened. Apparently on July 16, a faulty “configuration change” in the AWS billing system was implemented. “This system relies on unit conversion data to calculate line item charges,” AWS writes, but the change “caused updates to the unit conversion data to fail, resulting in inflated line item costs, which propagated to the Billing and Cost Management console and triggered the budget and cost anomaly alerts.”
Logs on the health dashboard show AWS trying to roll out a solution for about two days before marking the issue as fully resolved.
