As mentioned in the comments to the question, what you are looking for is "Feature flag management" aka "Feature toggle management", the terms are mixed but are the same. Feature flags do enable different patterns or activation strategies. Examples are among them you describe;
- "Kill switches", e.g. an easy way to disable a feature in production in case there are an issue
- "Release toggle", e.g. enable the feature after the code is deployed into production
Most Feature flag management systems will support the above and more activation strategies.
It is quite common to start by building such tools internally, essentially it is simple: Only a simple flag and a handful of if-statements. Your challenge is once you get dependent on your feature flag management. Usually you would need audit log capabilities, proper user management handling, proper scaling, ensure limited performance impact and similar capabilities. A nice introduction to feature toggle best practices to consider is found in this blog post.
There are multiple strong offerings available, these are the top four: