Feature Flagging: Proven Patterns for Control and Observability in Continuous Delivery

Are you moving faster than fast? Congrats! Chances are you already use feature flags to decouple code deployments from feature rollouts. Whether you use a roll-your-own feature-flagging solution (with a few quirks) or a feature-flags-as-a-service solution, hopefully you post feature flag rollout changes to something like Datadog so that when a flag rollout lines up with an obvious spike, you know whom to have a talk with. It's a solid start, but that’s just the basics. What do data-driven CD ninjas do? They build in observability to every feature release, so when they push a feature to 5 percent of users, it’s zero work to see how user and system behavior for those users compares to the other 95 percent. Ninjas also auto-calculate guardrail metrics, so when a change is pushed to locally optimize something, it’s easy to detect unintended consequences that impact organization-wide metrics. You’ll leave this session with a clear vision of how your team can achieve the same benefits, by either enhancing your in-house solution or adopting a commercial tool.