Background
I've set up vanilla "leaderboard" Meteor example. However, after ~12 hours of work, Meteor processes silently die (all of them, node* and mongo*).
There's nothing helpful in stdout, just the usual "App running at...". There's nothing suspicious in Kadira either. Memory consumption (~140 MB) and CPU utilization (~0.5%) are stable during whole process lifetime.
Question
Is there a way to figure out why Meteor processes were stopped? Have nodejs crashed? Have something sent a kill signal, or something else happend entirely? Right now I'm not even sure where to post the issue...
Setup details
Meteor is running in dev mode (insecure, etc) on Amazon EC2 t2-micro instance behind nginx. OS is Ubuntu 14.04 Server LTS. Connection upgrade to Websockets works correctly. Kadira instrumentation is enabled. Meteor version is 0.8.2