Best Practices & Workflow

Adopting feature flags is more than just a technical change; it’s a workflow change. Follow these best practices to keep your RocketFlag implementation clean and effective.

1. Naming Conventions

Consistency is key to finding flags as your project grows. Use clear, descriptive names:

• Good: enable-stripe-billing, v2-dashboard-rollout, maintenance-mode-banner
• Bad: test-flag, new-feature-1, jk-test

Pro Tip: Use Tags to categorise flags by team (team:payments), owner (owner:sarah), or release cycle (release:q3).

2. Flag Lifecycle & “Stale” Flags

Feature flags are intended to be temporary. Once a feature is 100% rolled out and verified, the flag should be removed from the code.

• Caretaker: On the Teams tier and above, RocketFlag’s Caretaker automatically marks flags as Stale (fully rolled out and unchanged for 30+ days) or Dormant (no evaluations for 60+ days), with a badge in the flags table. It can even generate a removal prompt for your AI coding agent. See Managing Stale Flags.
• The Workflow: Once a week, scan for Caretaker badges to find flags ready for retirement. Removing these flags reduces “conditional branches” in your code, simplifies your logic, and improves long-term maintainability.

3. Progressive Rollouts

Don’t just toggle a feature from 0% to 100%. Use the Traffic Percentage to mitigate risk:

1. 0%: Feature is in the code but disabled for everyone.
2. Always on for (Internal): Enable the flag for your internal team’s emails or IDs.
3. 10% Traffic: Release to a small subset of real users to monitor performance.
4. 50% -> 100%: Scale up as confidence grows.

Send a stable targetingKey (a user ID, not a per-session value) so rollouts are sticky: the same user keeps the same experience, and raising the percentage only ever adds users. See Sticky rollouts. To roll out to a slice of one segment, such as only Pro customers, pick an audience and set the percentage to apply within it. Keep attributes coarse (plan, region) rather than personal data.

4. Group Flags vs. Single Flags

• Use Group Flags for long-lived features that need to be tested in staging before production.
• Use Single Flags for quick experiments or simple applications that don’t have multiple environment stages.

5. Decouple Deployment from Release

The ultimate goal of RocketFlag is to allow you to deploy code anytime. By wrapping new features in flags, you can merge code to main and deploy it to production while it is still “disabled” via RocketFlag. This eliminates the need for long-lived “feature branches” and makes “Continuous Integration” much easier.