Establish logging pattern for non-critical failures

[wemeetagain]

@wemeetagain I want to establish this pattern that for non-critical issues we use the terms "failed" and "reason" instead of "error" as based on my observations the term "error" scares a lot of users, especially less technical ones.

Originally posted by @nflaig in #5285 (comment)

[philknows]

Also some backlogged things we've wanted to get to in regards to UX on #4400. In my opinion, generally, the word error: should be reserved for issues where a user misconfiguration or inaction by the user created a problem. Thus, it should only be used as a way to prompt users that they must do something to rectify it.

If it's not something a normal user can fix, it should be more of a warning type of message or something where stacktrace can be logged elsewhere but not saturate the CLI with endless errors a typical user cannot do much with.

[philknows]

Related to #5299 to set a standard

[philknows]

Closed in favour of #5359