All components are green, but production is down: analyzing an incident found in git

This article examines an instructive case of an incident where the company's infrastructure appeared fully operational according to monitoring metrics, yet production services remained unavailable for twenty minutes. Despite perfect database performance and no errors in the cloud infrastructure, the team could not quickly identify the cause of the failure. The author emphasizes that the problem was hidden in the infrastructure configuration change history, which engineers ignored during the first minutes of the outage. The analysis demonstrates that even with advanced monitoring and infrastructure-as-code, the human factor and reviewing Git history remain critical when troubleshooting. The article urges specialists to pay more attention to auditing infrastructure code changes, as they often contain non-obvious causes for major system failures that standard monitoring systems fail to capture.