Currently, the API returns an error when you try to do another action on a droplet while one is already running. For instance, if I use the API to snapshot a droplet and then immediately try to destroy the instance, the destroy request will fail. I understand the need to run these types of request sequentially, but it would be nice if I could ask the server to automatically queue the second request to run after the first. Currently, my application has to track failed requests and then retry them at a later time.
You could hide this functionality behind a flag which defaults to the current approach, but optionally allows queueing for those who want it.