Currently you need to reload the entire system, restarting all peers, routing and create a routing convergence if peers depended on the route. I suggest two things:
signal the peer if this side goes down gracefully (and reconnect using different timers, for a fast reconnect).
create childs and have a main proces that runs, kicking of chils for each plugin. Leave the main process (router) and childs (peerings) running unless something has changed. If there is a change either just add a new peering (its bad to have to restart everything to add something live) or only restart the affected peering child plugin.
This would create a lot more stable connector as if i am looking now, with 8 peers on a node, it takes quite some time. Also each peer seems to be started and connected in a sync way, causing it to take more time ... i dont see why this could not be done async as each peering should not be depending on each other and just collect routes.
Currently you need to reload the entire system, restarting all peers, routing and create a routing convergence if peers depended on the route. I suggest two things:
This would create a lot more stable connector as if i am looking now, with 8 peers on a node, it takes quite some time. Also each peer seems to be started and connected in a sync way, causing it to take more time ... i dont see why this could not be done async as each peering should not be depending on each other and just collect routes.