Provide functionality to write to disk if all forwarders are down

[tsheasha]

In case all forwarder are non-responsive, write the incoming messages to disk until the forwarders are back up.

The structure of the written messages will be one directory for each listener. Once the forwarders are back up, it will be the responsibility of relayd to to flush the messages from disk to the forwarders.