Health Columns for Feature Flags and Experiments

What

A new Health column on the feature flags and experiments lists surfaces the most severe signal on each flag, with a +N chip for additional issues and a hover popover listing all of them with environment context.

Why

Flags accumulate technical debt silently. A rollout left behind after an experiment ends, a rule referencing a missing experiment, or a value that fails schema validation can all affect production traffic without anyone noticing until something breaks.

How

• See flags and experiments that need attention at a glance. The Health column ranks issues by severity: High (invalid value, safe rollout should roll back), Medium (ramp needs approval, safe rollout unhealthy or without data, missing experiment reference, paused ramp, unreachable bar, rollout older than 30 days), and Low (temporary rollout).
• The Experiments list shows a parallel Health column for running experiments.

Learn more in the GrowthBook Stale Detection docs.