The most likely data loss scenario with S3 isn’t hardware/disk failure, but some software bug that corrupts data on ingress, or fails to persist and returns HTTP200. If you want to mitigate, do a read after write, and the odds of data loss become really minuscule.