Using Feature Flags to Make Data-Driven Decisions

LaunchDarkly is a powerful tool that allows you to easily accomplish safely releasing features, canary rollouts, targeting via our rules engine, and more. Today we will show you how you can improve the processes around these tasks by adding data. For example, if you are using LaunchDarkly to provide different variations of your product to different customers, you can improve that targeting by experimenting within those rules. Or if you’re doing a feature release or canary rollout, you can add metrics that allow you to validate the release is performing as expected so you can easily decide to roll back, stop the rollout, or continue releasing it.

Not only are we constantly changing our systems to match evolving needs, deep within our systems are an intricate web of dependencies which are in constant flux as well.

If you’ve ever had to change data sources, moving from a shared database to consuming an API, or switching from one database to another, you’ve seen this risk and maybe pain, first hand.We test extensively, everything works under test and we go live only to find something changed or broke when we went live, we missed something. Now we have to rollback or deploy hotfixes while production and more importantly our users suffer.

By launching darkly with feature flags, we can manage the release and adoption of new datasources deep within our systems safely and with ultra-granular control to avoid downtime, painful transitions, and safely validate in production before we go all in.

In this talk, you’ll get hands on with how to do this in the cloud with real world examples and walk away with source code you can take home.