Refactor logging levels to use multiple winston transports

[adamegyed]

Right now, there is only one winston transport, which sends all logs to a slack hook, and options in the config to adjust what is logged. This presents some added complexity which can be refactored out, and leaves the potential for certain logging events that could be useful in a postmortem analysis to be lost if the config is set to not include certain event types.

A solution to this would be to adjust the logging levels of the winston slackhook, possibly with custom levels, to only report on large-scale, relevant updates, while keeping the low level logging intact. We could add another winston transport that would save the logs to log files, which would give deeper insight to debugging without forcing the slackhook to receive massive amounts of relatively simple operational data.

Additionally, there are certain drawbacks to using slack for logging, such as a scrollback maximum based on the free plan. While it is convenient to view, having a fallback can be very useful in the event of data loss due to Slack message rollover.

[adamegyed]

This should also replace the usage of console.log anywhere in the code with winston logging at an appropriate level.