Automatically re-try processing files
Sometimes sounds fail processing for temporary reasons. For example, if processing server runs out of memory or things like that. In that case processing fails, but it could succeed if re-triggered after some minutes. I think it would be a good idea to add a new management command to be run every 60 minutes (or less), which gets all recently uploaded sounds that failed processing and tries to re-process them. For example, get all sounds uploaded in the last 5 hours that failed processing/analysis and send them to processing/analysis. If we run this once per hour, it means it will try to re-process 5 times and then give up.
There is. a command to handle that but it should be revised after the recent changes refactoring the analysis pipeline (which is different from the processing code but we could learn from it). See https://github.com/MTG/freesound/blob/master/sounds/management/commands/reprocess_failed_sounds.py
ideally we should try to re-process a sound X times, and if it fails, then we should notify users as described in the linked issue.