How should we apply log levels to log statements in the pbench-server?

[portante]

Proposal:

We expect the pbench-server to be normally deployed with the minimal logging level being warning.

The error and fatal log levels are to describe the operation of the server, when it has encountered something unexpected in its operation.

We'd emit warning log levels for items where some user operation occurred which does not affect the stability of the pbench-server itself, but the admin should be aware of what is happening (i.e. excessive login failures, excessive access failures to data which are not normal, very large requests, operations taking an unexpected long time, etc.)

info log levels are reported for most operational paths of the server reporting API behaviors of the user.

debug are low-level debug statements which dump extra data context for debugging.

trace are reserved for tracking entry/exit and branch flow of routines.

---

Related discussions from PRs:

• @portante in Implementation of user & access query semantics #2460 (comment)_
• @portante in Implementation of user & access query semantics #2460 (comment)_