Aaron Silvas

Looks like I identified the root cause. The `NodeHttpHandler` is waiting on a `continue` event on the request that never fires, and as a result times out after 1 second...

Not resolved, but with a hacky workaround I was at least able to demonstrate the real performance: ```sh bun tools/s3-test.ts Write testing myS3ExpressBucket... Read testing myS3ExpressBucket... Write rate: 1036.6 ops/sec,...

This bug has turned into quite the nightmare. The custom handler workaround that skips the `100-continue` and works for most cases, but hangs indefinitely when deployed into an AWS Lambda...

The linked PR might resolve this.

I've spent dozens of hours working around this issue but none have proven reliable across all AWS v3 SDK clients. I'm trying to avoid porting a major project back to...

Anyone waiting on a fix should thumbs up the original post to get it more priority.