General improvements to automation of data analyses, data processing, and other background tasks

This General Improvements issue is the high-level tracking issue for all the infrastructure upgrades. Kinda. These issues are getting messy.

We are encountering a need for better automation of particularly forest tasks, but there architectural issues we need to address.

providing data on the website about live data processing server status
improving the celery broker behavior - we literally can't reliably get a list of running tasks right now, even on the data processing servers. There are several possible reasons, chief suspect being rabbitmq just kinda not being sufficiently good or fast.
~We need to update celery. In general we also want to have pipelined stages of processing that tasks run with allowing us to display detailed data. (nothing beats using your own custom tracking)~ Done. needs to go out on servers other than staging, but done.
The celery time-values-under-6-minutes-just-don't-work bug may have been a pytz timezone offset bug! more frequent top-up plus a reliable broker so we can see queued tasks may be the basic improvement we need.
a refactor of [uploaded] data processing is absolutely required in order to extend it and provide statistics. It is too many deeply nested functions. Effort has been made on some sections of the code by creating various classes, this needs to be picked up and continued.
We now have a ~generic background task runner, we should integrate that to allow site admins to dispatch long-running tasks, particularly a data deletion (and eventually compression) feature. (data deletion is in development as we get various compliance features/requirements added to beiwe.)
(future wish-list: shift to more of a container system)

(heyyy it displays them nicely inside bulleted lists)

onnela-lab / beiwe-backend