I’m looking to build a robust exception/error logging mechanism. This is all in-house, meaning for all my client’s apps. The plan is, when there’s an exception in one of the client applications (there are many clients), that client will “push” this out to a centralized “hub” to a REST endpoint (build in Meteor) where I’ll collect this data for reviewing/reporting/alerts etc.
I have it all set up and right now log the error.name, error.message, error.stack, userId, appid and a few breadcrum tracers.
What would you log, keeping in mind reporting/viewing scenarios?
This is one of the things Kadira does btw, both on the client and server. Though if you really only want error reporting and not performance metrics, building your own simpler system sounds good.
Maybe you should try setting up Kadira though. It’s a bit fiddly, but it’s pretty useful, and when compared to building your own system it shouldn’t take too much time.
Working on a project now using enhanced logging with Winston. I’m logging just about every single route and major function at the silly log level, which is very handy dealing with bugs in development.
Otherwise just basic info about user accounts, method returns, functions that can throw and other async stuff at the debug level, and all errors at the error level.
We don’t have a lot of users though, so we can easily handle a lot of logging