also failed requests are now sent to sentry

* using gevnts for async requests handling of gunicorn
* serving the static frontend via gunicorn
* split requirements into dev and general
{