It's just version 3.x of the AMI that is being discontinued on AWS - version 4.x is still available within their Marketplace (and on Azure too). So, you just have to upgrade to stay on Amazon if you're using the old flash version.
I would suspect that the behavior would be the same if the progress callback was timing out during an HTTP transfer. Every so often when transferring files we call the callback and timeouts take time.
The problem seems to be coming from a timeout when requesting your callback. Since it's just a progress callback, the error does not stop the job, but it does take a minute for the request to actually time out. Suggest you get to the root of the problem it's having requesting the callback.
There is no way to abort a running job, short of killing the moviemasher.rb process and manually deleting the SQS message.
Sounds like one of the path related options is either not set or set to somthing that starts or ends with a slash.
Transfers of type file are just local files - files on the same machine as moviemasher.rb is running. I'm not sure what you mean by imageconversion files. The files being transfered should be in the destination directory if the transfer succeeds.
That is still the case, yes
The system could either have crashed entirely, or had trouble requesting the callback endpoints. With debug logging enabled on the transcoder the difference should be discernable.