To those people who are worried about database. Let me share what i think about that...

Suppose your database has 1 million users. You think your app will be in troubles ?

The main idea of "feature-flags" is NOT EVERY user will have access to that feature. So, if you are doing really an experiment you will choose may be the 1% of that users... 10.000 users for this case. That number is a joke for any database today.
When you realize that "feature-flag" experiment has time and objectives to measure after the experiment ends... you will be not using 1000 feature flags at the same time for the 100% of users. May be 2 or 3. So, 10000 users * 3 FF = 30000 rows! It's still a joke for any kind of database.
May be if you think about feature flag not like "if / else" directive, you will understand better the real usage, purpose and benefits of using this. And probably you will not have performance issues or troubles because of this.
This is a package provided from Laravel team, so this is a start point. Surely in a couple of months some developer will publish a package with improvements, like Spatie has made with a lot of features of Laravel.
This is a programming technique, so... if you are a programmer and not just a "framework configurer", you can overwrite this through OOP, like inheritance, and build a new functionality which may be could simplify the caching used by default. In example, if you want just 1 row caching all the feature flags for same user, you can use a single row which stores: id, modelName, features. Where features columns is a JSON column having an array of objects with {"feature" : "dashboard_v2", "value" : true}. Even today you can query JSON through SQL syntax.
May be you are thinking in a real case which you use and you know that database works slowly or is full of bottlenecks and that's what really bothers you. Think that the problem in that case is not this new feature, the real problem is another thing that you will fix with another solutions.

So, there are a lot of solutions from this start point. Find solutions, or problems to solve, that's programming. Thanks Jeffrey for this explanation, and Laravel Team for this much needed feature.

Laravel Pennant