In the previous newsletter (as well as in my website article) , I covered trunk-based development. One of the ways to keep branches small is to commit and and merge to the main branch often, by keeping part of the codebase in main inactive. Feature flags is a way to do it systematically.

Feature flags, also known as feature toggles or feature switches, are a powerful technique used in software development and data science to enable or disable specific features of a system at runtime without requiring a code deployment. They provide developers and data scientists with greater flexibility and control, leading to several key benefits:

  1. Incremental deployment and testing: Feature flags allow teams to separate code deployment from feature releases. This means that a feature can be developed, tested, and deployed incrementally without affecting end users. This approach leads to faster iterations, reduced risk, and a more efficient feedback loop.

  2. Controlled rollout: With feature flags, new features can be gradually rolled out to a specific subset of users or servers, enabling controlled experimentation and analysis. This selective exposure helps to minimize potential negative impacts and allows for A/B testing, performance monitoring, and user feedback analysis.

  3. Safe failover and rollback: If a new feature causes issues or negatively impacts the user experience, feature flags make it easy to disable the problematic feature without a full rollback or redeployment. This reduces downtime and mitigates the risk associated with deploying new features.

  4. Customization and personalization: Feature flags can be used to provide tailored experiences for different user segments, allowing teams to deliver personalized features and content to specific groups of users, such as premium subscribers or beta testers.

Here are the two simple code snippets to understand how feature flags work. First covers configuration of feature flags and second activation of the feature based on feature flag. There may be more sophisticated implementation of this paradigm, including Inversion of Control, I just wanted to provide a simple intuition behind it.

feature_flags = {
    "my_feature": True,
    "another_feature": True
}

if feature_flags["my_feature"]:
    activate_my_feature()
if feature_flags["another_feature"]:
    activate_another_feature()

I’d encourage you to try feature flags in your next project and experience the control and flexibility that they provide.

If you like what I write, consider subscribing to my newsletter, where I share weekly practical AI tips, write my about thoughts on AI and experiments.


This article reflects my personal views and opinions only, which may be different from the companies and employers that I am associated with.