lack of queue size management could lead to memory exhaustion, crash.

[markstos]

It appears that if the Syslog server is unavailable via TCP, the message is pushed on to an internal in-memory queue in the form of the self.queue JavaScript array.

If the syslog server continues to be down and logging continues to arrive, there is nothing to prevent the size of this queue from growing until memory is exhausted and things start to crash.