Nope. No S3 logic behind the scenes. A few of the ELBs are fine, and a few are not. Seems random.

The EC2 instances themselves are fine, but the affected ELBs are spitting out 500s.