RFC[Ouroborus]: deploy other cotainers, allow orchestration

[jeffque]

As of now, Ouroborus is capable of respawning itself (calling hestia).

We must do a way that allows Ouroborus to spawn new containers (daemon mode), so that it can create some level of orchestration.

• Must be non-opinionated as possible (one constraint is that a container must have name in ouroborus) #23
• It shall not imediately start a server (as opposed from current version that calling a function make it start) #24
• Object that describe desired container #25
• Option to destroy/restart container #26

bin/ouroborus will contrast with ouroborus as a lib.

• it will start a container, and also only accept a limited set of predetermined containers #19
• depending on the args passed (or env var), it will discover which containers it shall run, each of then described by a ruby file #20
• it will keep a track of containers upped in a mongo DB, which will be it's source of truth #21
• [exec] recover state #22

[jeffque]

Some containers depends on networks. I believe that networks could be another resource to be allocated (as are containers per #25 ).

Could differentiate containers in apps and services? Also, networks should be a thing that containers could rely on. And go live to a container would guarantee that the network is also up