When the CI Broker runs and the Radicle node stops or restarts, the broker will fall into an endless loop printing these logs :

 DEBUG ci_broker                 > waiting for event from node
 INFO  radicle_ci_broker::event  > no more events from node
 DEBUG ci_broker                 > waiting for event from node
 INFO  radicle_ci_broker::event  > no more events from node
 DEBUG ci_broker                 > waiting for event from node
 INFO  radicle_ci_broker::event  > no more events from node

until I somehow terminate it.

it would make sense for the broker to terminate itself when this happens. Maybe, it could check if the Radicle node comes up so that it can connect to the socket and work again normally.