Adding/Changing binary types on APIGateway creates temporary downtime

I tried adding support for binary types on my API but have been facing an issue where APIGateway seemingly breaks for a period of time after adding/changing the supported binary types. I’m talking 1hr+ and then it magically starts working again. It doesn’t even get to invoking the lambda, just returns 500 with an empty response. Has anyone seen this happen before? I don’t know if this has something to do with serverless or if it’s an AWS bug.

Any help would be great. I’m concerned about advancing this deployment to production incase it’s not a one-off.

Thanks.