CF workers breaking with AWS S3 starting June 1st

Starting June 1st all requests to has started receiving a 403 error from our AWS S3 instance. We changed nothing on both AWS or CF. The content is still publicly accessible when going directly to S3 from my laptop and mobile. Anyone else face this issue?

Here is the CF link
Here is the same file from our S3 bucket :

Looks like you resolved this in the community Discord

The pathname from the URL constructed from request.url starts with a / but so did your hostname, so your resulting path started with // which AWS refused - for completeness or anyone else who is curious.

1 Like