The Meteor tool process would run multiple big "parts" that should be managed: started, stopped, restarted, monitored for crashes, etc.
The interface that Meteor tool uses is called a "runner".
Some of the runners run sub-processes. The main runner is called run-all, it
has sub-runners such as run-app, run-mongo and run-proxy. Other runners
are used for different CLI commands.
The output of the runners (compilation errors, state change, etc), should be reported through the Run Log. Run Log is also something that users see in their browsers, when their app breaks.