What is a feature flag?

A feature flag is a configuration switch that controls whether a piece of functionality is enabled, letting teams turn features on or off without deploying new code.

How is a feature flag different from a canary deployment?

A feature flag controls visibility of a specific piece of functionality at the application level, while canary deployment controls how much traffic reaches a new version of the application at the infrastructure level; the two are often used together.

Why use feature flags instead of just deploying when a feature is ready?

Feature flags let teams merge and deploy code continuously while controlling exactly when a feature becomes visible, reducing the risk and coordination overhead of large, infrequent releases.

Flag debt refers to the accumulation of old, unused feature flags left in a codebase after a feature has been fully released or abandoned, which adds unnecessary complexity if not cleaned up.

Can feature flags be used for A/B testing?

Yes, feature flags are commonly used to expose different variants of a feature to different user segments and measure the impact on key metrics.

DevOps

Feature Flag

BeginnerConcept1.2K learners

A feature flag (also called a feature toggle) is a configuration mechanism that lets teams turn a piece of functionality on or off at runtime, without deploying new code, decoupling code deployment from feature release.

#FeatureFlag #DevOps#Concept#Beginner#CanaryDeployment #RollingDeployment #BlueGreenDeployment #Harness #SoftwareDelivery#Glossary#SkillVeris

Definition

A feature flag (also called a feature toggle) is a configuration mechanism that lets teams turn a piece of functionality on or off at runtime, without deploying new code, decoupling code deployment from feature release.

Overview

Feature flags wrap a piece of functionality in a conditional check against a remote or local configuration value, so that a feature can be shipped to production in a disabled state and then enabled later — for everyone, for a specific percentage of users, or for a targeted segment — without a new deployment. This separation of "deploy" from "release" lets teams merge and ship code continuously while still controlling exactly when and to whom a feature becomes visible. Flags are commonly used to implement gradual rollouts similar to canary deployment, to run A/B tests comparing feature variants, and as operational kill switches that can instantly disable a problematic feature without requiring a rollback or redeploy. Dedicated feature-flag services and platforms — including capabilities built into broader delivery platforms like Harness — provide targeting rules, audit logs, and real-time flag updates so changes propagate to running applications without a restart. Because flags accumulate over time, teams need discipline to remove stale flags once a feature is fully released or abandoned; otherwise the codebase accumulates conditional branches that add complexity and risk. Feature flags are often used alongside deployment strategies such as rolling deployment or blue-green deployment, giving teams two independent levers — infrastructure-level traffic routing and application-level flags — for controlling risk during a release.

Key Concepts

Runtime toggles that decouple code deployment from feature release
Percentage-based or segment-based rollout targeting
Kill-switch capability to instantly disable a problematic feature
Support for A/B testing and experimentation
Real-time flag updates without redeploying or restarting the application
Audit logging of who changed which flag and when

Use Cases

Gradually rolling out a new feature to an increasing percentage of users

Running A/B tests to compare feature variants

Acting as an operational kill switch to disable a broken feature instantly

Enabling early access to features for internal teams or beta users

Shipping incomplete code to production safely behind a disabled flag

Frequently Asked Questions

From the Blog

Data Science

Scikit-Learn for Beginners: Machine Learning in Python

Scikit-learn is the most widely used Python library for classical machine learning. This guide covers the fit-predict workflow, train/test splits, classification, regression, model evaluation, feature engineering, and pipelines — everything you need to build and evaluate your first ML models.