Jan Giacomelli • 6/22/2026

How I destroyed production OpenSearch cluster

This article details a real-world incident where a production OpenSearch cluster was accidentally destroyed due to a series of mistakes. The author explains how environment drift, outdated Terraform configurations, lack of plan review, and overly permissive IAM roles led to a destructive plan being executed. The post highlights the importance of reviewing infrastructure changes, keeping environments in sync, and implementing proper guardrails. It serves as a cautionary tale for DevOps and software engineers about the cascading effects of small oversights in cloud infrastructure management.

0 comments

#DevOps #Infrastructure As Code #Terraform