Toggle is a word that describes switching between two states — for example, toggling between screens as you video chat with two friends at the same time. This simple user-interface element is also used in software and hardware to switch between different functions, settings, or options. It’s important to use clear labels and visual cues when using toggle switches to make it easy for users to understand what each state means and the consequences of their actions.
Toggles that enable new features should be tested before they go live in production. This includes testing with all toggles flipped On (as that will be the production state) and also with all toggles flipped Off. The latter will be the fallback state and is recommended for testing to avoid regressions in future releases. Additionally, it’s important to test with all toggles flipped Off for any existing or legacy functionality that might be impacted by the new feature toggle configuration.
Managing toggle configuration can become difficult at scale. Many teams choose to move from a static file based approach to a more centralized system like an application DB or a feature flags registry. This allows for more dynamic in-memory re-configuration and is often paired with some form of admin UI that product managers, testers and other stakeholders can access.
If you’re not already doing so it’s a good idea to adopt an API-based toggle management system. This makes it possible to perform experimentation with your toggles at runtime and is often paired with some form of CI/CD integration that ensures your re-configurations are always applied in a consistent manner across your entire production fleet.
Another option for managing toggle configuration is to hardcode it into your code. This is usually done through some type of conditional statement like a #ifdef or preprocessor directive. While this is a viable method to manage toggles at scale it can be cumbersome and slow. Often times it will require restarting your process or at least re-deploying an artifact in order to change the toggle configuration. This can significantly slow down your validation cycles and impede the speed at which you’re able to respond to feedback from your customers. Additionally it may not be as secure as more asynchronous toggle management methods which use an in-memory runtime to control configuration.
