A one-line Kubernetes fix that saved 600 hours a year

This engineering post describes diagnosing and resolving very slow Atlantis (Terraform orchestration) pod restarts caused by Kubernetes recursively changing file ownership on a persistent volume with millions of files. By setting pod.spec.securityContext.fsGroupChangePolicy to `OnRootMismatch` (instead of the default `Always`), the team reduced restart time from ~30 minutes to ~30 seconds, reclaiming approximately 50 hours of blocked engineering time per month.