Example of using feature flag commands with gitlab-experiment

What does this MR do?

This is an example of how to hook into the lower level feature flag layer by using values within an experiment context, instead of using the experiment context.

It can be one way to implement a gitlab-experiment in such a way that you can utilize the feature flag chatops commands, but only takes into account a minimal experiment definition in that multivariate logic would not be possible in this experiment anymore.

As is outlined here, we have to define the custom experiment class -- which we do here, because we already have a null_hypothesis experiment with a matching feature flag defined. The null_hypothesis experiment has been running for a while already, and this would change the "percent of actor behavior", since we're changing the underlying actor ids. Read more about Flipper Gates if you need more information here.

I don't plan to merge this MR, but if someone does want it merged, the will need to communicate that it would distribution percentages from 25/75 to something else (which is not reliably predictable).

You can ignore the first commit -- it's an unrelated fix that !55792 (merged) resolves.

Does this MR meet the acceptance criteria?

Conformity

• 📋 Does this MR need a changelog?
  • I have included a changelog entry.
  • I have not included a changelog entry because _____.
• Documentation (if required)
• Code review guidelines
• Merge request performance guidelines
• Style guides
• Database guides
• Separation of EE specific content