NodeJS doesn't have a built-in logging framework. The standard beloved JS
console.log can be used, but there are several (competing) 3rd party solutions that work pretty well.
The demo project uses
winston package. As a package must have a name,
winston can be just as good of a name for a logger as, e.g.,
In the past, configuration around logs has been an important component of any server software install process. Logs would be rotated, archived, passed on to a monitoring solution.
winston has a complete set of capabilities configuring log outputs.
However, containerization may be the death of old-stile loggers. Basically, the console log output is all we need when running a server in a container, for either dev or production. As Docker reads the console-out, it seamlessly passes it on to the monitors and/or consolidators, bypassing anything that has to do with dedicated log files and their rotation.
Technically, docker-gathered logs should be rotated on the host. Not in scope of this course.
So, in the demo project, the use of
winston is limited to enhancing the console output, and the purpose of expanding beyond the core JS
console.log is strictly to get better formatting capabilities. If you wonder why
file: configuration is in
src/utils/logger.js - it came with the usage sample. Feel free to delete it.
Finally, before we get to running live queries and mutations, a necessary walkthrough to see how to get GraphiQL to work - our Front End tool. A bit convoluted, but not too bad!