Introducing a stage environment

[Makman2]

Have corobo changes land in a staging instance before deploying to production (what keys will be different in staging? If none, this is useless. If some, the staging wont be testable unless labhub is connected to a different ‘dummy’ org, and then nobody will use it.)

Separated from #509

[Makman2]

Currently I don't see major benefits in introducing a staging system compared to the effort to introduce one, because

• someone has to conduct manual tests on it to make it actually useful. I rather prefer to do continuous deployments and fix things fast. Problems are then covered by automated tests instead of manual ones (leads to more reliable deployments).
• necessity to manage infrastructural resources needed by the staging system (an org, gitter channels, secrets, ...)

The only benefit to me is

• allow to quickly test new features by deploying to stage.