I recently went through implementing a home-grown feature flagging system. It made me realize the complexity and trade-offs one must decide with feature flags. I think this is an area where a lot of teams may tackle and use feature flags differently. It’d be great to get some shared knowledge out into the world.
- How does your team handle feature flagging?
- What do you use them for? (release, pricing plan features, etc)
- Do you provide a different system for other team members (ex. support/marketing)?
My current company uses configs for feature flags, right now that still requires us to redeploy the app to toggle features (not the most ideal).
My previous one used Flipper with Redis (Rails), and that worked really well for us. I think FunWithFlags in Elixirland does the same thing, that’s probably what I’d pick up in my personal projects
Did it in some small project. Database table with key and boolean flag status, data loads from db to ets on boot and when someone toggles flag via UI. Works well. Project scale was too small to talk about teams/access/usage practices though