Drasko DRASKOVIC

@manuio I'll review

This must wait for @dborovcanin's approval

@fthdrmzzz @manuio when can this one be finished?

This does not looks to me like a problem that we should address in the code, but it should be rather addressed byt deployment itself (either via K8s scrips or...

@MisterScience5 I would accept this PR that @nmarcetic proposed, but not pollute the current `docker-compose.yml`. Our (Mainflux team) preferred way of deploying in production is Kubernetes, `docker-compose` is used mostly...

Makes sense, but I think: - It should be really trivial, so that it merits lean implementation rather than dragging whole can of worms (I see tools like `kafka2nats`, `stan2kafka`,...

This can be easily done through [mProxy](https://github.com/mainflux/mproxy). Closing here, as this will not be main focus of the Core.

@chombium @nmarcetic any opinions?

Agreed. Only thing that I would really like to see is a kind of abstraction around Things and Users microservices DB (so - not messaging, but metadata DBs) to replace...

Agreed - [Viper](https://github.com/spf13/viper) looks like a nice tool for the job.