Rqd bindins to a single cuebot instance until restarts

[DiegoTavares]

Describe the bug This is a scalability issue. Whenever a new cuebot is added to the pool, one of two things need to happen to have existing rqds to connect to it: All rqds need to be restarted, or all existing cuebots need to be restarted.

To Reproduce Steps to reproduce the behavior:

1. Start 1 cuebot instance and 3 rqd instances
2. Check all rqds are connected to the same cuebot
3. Start another instance of the same cuebot (accessible through the same hostname using a load balancer)
4. Check that all rqds are still connected to the same cuebot instance

Expected behavior Ideally rqds should refresh their connection channel from time to time to make sure load is always distributed between the available cuebot services.