Building your own feature service? Yeah-Nah!
Updated: Oct 18, 2020
In the last several years I have been working with many clients who have struggled to decide which feature flagging platform to choose. Often it would come down to cost and an unwillingness to take on additional expense. It seems as though it has been difficult to quantify the value of feature flags and experimentation in production to those who make overarching decisions, possibly because this technology is still not widely used and without the buy-in required to embed it as standard practice. As a result, many teams I have worked with built their own feature flagging service, with such a decision coming from the developers themselves. It is a great thing when you can build a service and have full control of it, however what a lot of teams don’t realise is that feature flags are quite an intricate technology, it needs to evolve with time, and like all products requires constant enhancement and maintenance. Therefore, the software teams end up building a very simple solution that doesn’t necessarily cater for the entire team. For example, it could be a feature service that doesn’t have an interface, so your “business folks” could never turn a feature on or off in production, which really defeats the purpose of the entire approach. Another problem is that often it doesn’t have important functions, like gradual rollouts and analytics in order to perform more complex A/B scenarios, let alone enterprise level security support. In addition to that, if your enterprise uses multiple tech and programming languages, it is likely they will be building multiple feature flagging services as opposed to having one universal service for the entire company.
With FeatureHub you don't have to worry anymore about such problems. It includes all the basic functionality of a feature service plus more advanced features, like access level permissions, authorisation through Gmail, variety of SDKs, and support for client, server and mobile apps. It is absolutely free and open source, so grab our Docker image, install on your own infrastructure and try it out with your team.