Assess the usefulness of feature flag state change logging strategies
We currently have two ways to log our feature flag state changes in our environments:
- Internal GitLab.com feature flag state change issues: For each change of a feature flag state on GitLab.com, an issue is created in this project. This is cross-linked to rollout issues, e.g. gitlab-org/gitlab#415503 (comment 1598380833).
-
Internal GitLab.com feature flag state change logs: Filter logs with
source: featureandenv: gprdto see feature flag state change logs. This is linked from Production incidents under the "📚 References and helpful links" section, e.g. gitlab-com/gl-infra/production#16366 (closed).
The way I see it, both strategies have their own advantages and audience, but I'd like to get actual feedback from the people that are actually using them (or not).