feat: Overhaul logging behavior

[Histalek]

Current behavior

Gatus spams the journal with errors (priority=3, log-level=error) as every message from Gatus is logged the same way.

I'm running Gatus in a container with podman and the journald logging driver.

Expected behavior

Log messages from Gatus are logged with an appropriate log-level. Not spamming the journal with error messages that are essentially no errors.

Proposition

1. Log output to stdout and log diagnostic information about Gatus itself to stderr
2. Give every log message proper log-levels

Reasoning and other comments

The current behavior is most likely caused by logging everything to stderr without explicitly setting a log-level/priority (log.go does this by default).

IMO stderr should be used for diagnostic information, e.g. listening address, database connection, configuration changes and validation, anything needed to possibly oversee the operation of Gatus itself.
Watchdog messages should then instead be logged to stdout, because these messages are more likely the output of Gatus and should be handled as such.
Especially for watchdog messages i'd like to see INFO for successful checks and WARN for unsuccessful checks as the log-level.

I'm not sure if all of this is possible with log.go, maybe something like logrus, zerolog or zap would be more suitable.

[TwiN]

We could just resolve this by setting log.SetOutput(os.Stdout). By default, log uses os.Stderr.

[jangrewe]

@TwiN To chime in on this, i'd really like to disable the output of regular checks, as >200 checks creates quite a bit of log output. ;-)

[TwiN]

@jangrewe @Histalek this has been resolved by #872 and #895. Not released yet, but it's available under the latest tag.

A bit late, but better late than never.