Is your feature request related to a problem? Please describe.

I would like to be able to write lambda functions in Go that process and return more data than can easily fit in RAM, e.g. transformations on / aggregations of large S3 objects.

Describe the solution you'd like

An io.Reader and/or io.Writer-based Handler alternative that allows me to write to a stream rather than fit my response in a byte slice.

Describe alternatives you've considered

1. I can write the data to intermediate storage in an S3 bucket and return the location of the S3 object, but that adds overhead and complexity, and incurs storage and access costs.
2. I can write a web app instead of a lambda and deploy it in a container using Elastic Beanstalk, but that adds various kinds of complexity I was hoping to avoid.
3. I can use a different language, such as Java or C#, that supports streams, but if I was happy programming in Java or C# I wouldn't be writing Go programs.