Currently it only uploads one log line with each HTTP request, which does not work well for high velocity logs.

The queued handler could potentially read all the available lines in the queue and upload everything outstanding at once as opposed to popping an item one at a time, and uploading it