It's important that we have `go.sum` `go.mod` and `vendor` directory in
a synced state. It's error prone having each developer checking this
manually every time they change a dependency. Create a new job that will
check this to make sure everything is in sync.

Example of failed pipeline: https://gitlab.com/gitlab-org/labkit/-/jobs/429282504

Unlike some of the other functionality provided by Logkit,
the logging functionality does not wrap an underlying log provider.

Why not? Because, in a past life, I worked as a Java developer and
therefore understand that log wrappers will bring you nothing but pain
and unhappiness.

Instead of being a log wrapper, this functionality encourages
applications to use logrus. Logrus is already widely used in GitLab.

What benefit does this functionality add:
* Consistent configuration of logrus
* Importantly, this functionality allows labkit itself to write logs.
  * This will allow us to add better logging, as well as standardised
    access loggers, exception loggers etc

Co-authored-by: Zeger-Jan van de Weg <zegerjan@gitlab.com>