Release Engineering: Scenario-Based Questions

Feature flags allow you to ship code continuously but release features selectively. They’re key to controlled rollouts, A/B testing, and safe experimentation — when implemented thoughtfully.

🚩 Types of Feature Flags

• Release Flags: Turn features on/off post-deploy.
• Experiment Flags: Enable A/B testing or canary cohorts.
• Ops Flags: Temporarily disable behavior during incidents.

🧱 Flag Implementation Options

• Boolean flags: Simple enable/disable.
• Targeting rules: Enable for specific users, groups, or conditions.
• Dynamic rollout: Gradually increase exposure by % or region.

🛠️ Tools & Services

• LaunchDarkly, Unleash, Flipt
• Homegrown flag services backed by Redis or config stores
• CI/CD plugins for toggling flags during deploy

✅ Best Practices

• Externalize flags (no redeploy needed to change state).
• Audit flag changes and restrict write access.
• Set TTL or cleanup plans for temporary flags.
• Use flag status metrics to detect partial failures or crashes.

🚫 Common Pitfalls

• Leaving stale flags — creating tech debt and complexity.
• Toggling flags without validating downstream dependencies.
• Flag logic deep inside business logic — hard to test or remove later.

📌 Final Insight

Feature flags turn code into controlled experiments. With safe toggling, metrics, and governance, they unlock continuous delivery without compromising safety.