Transient broken connection with RabbitMQ during deployment

[Datamance]

TL;DR:
Coordinate deployments in such a way as to avoid this kind of transient error:

consumer: Cannot connect to amqp://admin:**@production-lookit-rabbitmq:5672//: ConnectionRefusedError(111, 'Connection refused').
Trying again in 32.00 seconds... (16/100)

Narrative
As a developer, I want my alert channel to convey as much information as possible. When there is too much noise from deterministically produced errors related to deployment, the channel becomes less useful.

Implementation Notes
Solution will probably solve #1 and #5 simultaneously as well, given the fact that they're both transient deployment issues.