Incident: An agent's cleanup destroyed production
Module 9 · Incident file · About 35 minutes
What happened
On 6 March 2026 Alexey Grigorev, founder of DataTalks.Club, published "How I Dropped Our Production Database and Now Pay 10% More for AWS" on his newsletter, Alexey On Data. He was moving his AI Shipping Labs website to AWS with Claude Code, and "I was overly reliant on my Claude Code agent, which accidentally wiped all production infrastructure for the DataTalks.Club course management platform". That platform held "2.5 years of all submissions: homework, projects, leaderboard entries".
His timeline starts on Thursday 26 February at about 10 PM, deploying with Terraform without the state file, "as it was on my old computer". At about 11 PM a Terraform auto-approve command "inadvertently wiped out all production infrastructure, including the Amazon Relational Database Service (RDS)", and "all snapshots were also deleted". He upgraded to AWS Business Support at midnight, which "costs me an extra 10%"; AWS found a snapshot on its side, and by about 10 PM on Friday the database was back, with "1,943,200 rows in the courses_answer table alone". He writes: "This incident was my fault". Everything here is from his post.