log levels

[travi]

Is your feature request related to a problem or a specific use case? Please describe.

i've wanted to include the more detailed logs from traverson in our log aggregator, but hadn't had enough reason to follow through until today. i was able to pipe from minilog to a writable stream that turns around and calls logger.trace() on our instance of a bunyan logger.

as i looked deeper into the traverson log output, i realized that there are some of the details that provide a useful story of the steps that happen, but others that are probably a little more detailed than would be useful in this case. i see that minilog provides some level of filtering that could be applied before i pipe to bunyan. however, it appears that all of the traverson logs are at the debug level.

Describe the solution you'd like

more separation of levels would enable simple whitelist/blacklist filtering by those levels. do you think there would be some level of logical grouping that could be applied to the logging?

the logs that seemed natural for me to aggregate were the ones like:

• link for x found
• request started
• request returned

logs like the implementation of a transform function to be applied seemed less useful for the use case i was after, so could possibly make sense at a more detailed log level.

Describe alternatives you've considered

since minilog and the writable stream is all pipes, i could certainly write my own custom filter for minilog or be more selective of what was actually written to logger.trace() on the bunyan side of things. however, those options would require deeper inspection of the content of the messages and would run the risk of not matching if that content changed in the future.