fix(runners): increase the log level to WARN when using the enable_job_queued_check parameter

[Fgerthoffert]

The readme section detailing the enable_job_queued_check makes it an appealing function, but it was unclear to me that it could cause some issues when used with ephemeral runners and matrix workflows.

It is particularly challenging to notice for organizations with a very large number of repositories, at times with low activity the issue is not present, but when matrix jobs are started or when multiple repositories are triggering jobs around the same time window (such as cron'd workflow), this issue is more likely to be faced.

For example, on the execution that allowed me to find the cause, when starting a matrix of 9 jobs, 4 of them were actually triggering this log message:

2023-03-10 15:29:09.728  INFO  [scale-up:a0502c68-f9b9-5d2d-afd6-70edb7486316 index.js:135249  isJobQueued] Job not queued 

This log message, aside from being one amongst many other INFO messages, does not provide many details about the consequences.

I suggest bumping the level to WARN and adding more details in the message itself.

Related discussion: #2931

[npalm]

LGTM

[npalm]

I think the logging messages we have can be improved. Thanks for your contribution. I have also created a PR to replace the current framework (which is out-dated). The replacement will not direct change the log messages. But provides with each messages the context. And we drop support for line based logging, moving to json which is much easier to parse. See #3037