Include top-level Sidekiq warning in Dangerbot output
This issue is a spin-off of the discussion in #408911 (closed) (Provide process to avoid sidekiq problems with mixed deploys).
In order to help avoid situations where the engineering guidelines around changing sidekiq arguments are missed, we could expand Danger Bot's behaviour to:
- Include a top-level warning in the danger-bot output to link to the guidelines.
- In both the top-level and inline comment, include a brief description of the importance of the guideline and the potential impact of not following it.