Describe what means 'definition of done' for Regressions and Bugs

What does this MR do?

I looked at our Definition of Done and I improved the definition of fixing bugs and regressions, as this is not explicitly stated.

We should always aim to fix issues in a way that does not only solve the underlying problem, but also extensively covers that with relevant tests that such will ensure that such issue hopefully will not happen again.

This might not be clear for contributors that this is the most important aspect of fixing regressions / bugs. This makes it clear.

Does this MR meet the acceptance criteria?

Conformity

• Documentation created/updated or follow-up review issue created
• Code review guidelines

/cc @gitlab-org/maintainers/rails-backend