Why no upload?
Opened this issue · 2 comments
duncanphillips commented
Why no upload?
pfdevilliers commented
A possible way to implement this is with go routines. We perform the upload to S3 within a go routine and then return immediately to the caller. We can integrate this with something like SNS to notify the caller when the object has successfully been uploaded OR add in a polling API mechanism to check the progress of the upload.
This really depends on the size of the file being uploaded (or the number of simultaneous uploads).
qstorey commented
Yes definitely, uploads should be handled as quickly as possible by storing the content to disk and returning a 200 on the POST request.
By handling HEAD requests (currently not implement), the caller can then make subsequent HEAD requests to determine if the upload to S3 has completed.