Feature Flags

Release when you want. Control what users see.

Create, target, and manage feature flags for progressive delivery at enterprise scale. Accelerate releases without harming the integrity of your application or your customer experience.

Why Feature Flags

Release with control, not hope.

Govern every change, target exactly who sees it, and release gradually with the confidence to roll back instantly.

Govern

Every flag change is policy-checked, pipeline-gated, and approved before it releases, no exceptions.

Target

Target any user or segment on any attribute you define. No hardcoded logic, no redeploys.

Release

Release to 1%, watch your guardrail metrics, and expand only when the data says it's safe. One click kills the flag if something breaks.

Govern

Flags with policies, pipelines, and approvals.

Harness treats flag changes like any other change in your delivery pipeline, subject to the same OPA policies, approval chains, and audit requirements as your code. Governance isn't bolted on; it's built into every release.

OPA policy enforcement on every flag change. Write policies once in Open Policy Agent and apply them across flags, configs, and pipelines. Block non-compliant changes before they reach production, automatically.

Pipeline-native approvals via Harness Pipelines. Flag activations trigger inside Harness Pipelines alongside your deployments. Approval gates, JIRA ticket validation, and environment-specific rules all in one place.

Coordinate complex releases with Rollout Boards. Give large teams a single view of every flag status and outstanding approvals.

Target

Every feature targeted to the right user.

Target any user or segment based on any attribute you define. No hardcoded logic, no redeploying, just precise control over who sees what, in real time.

Target by user, segment, or any attribute. Serve features to specific IDs, cohorts, geographies, or custom attributes with fine-grained rules.

Ramp from 1% to 100% on your terms. Start with internal users, expand to beta cohorts, then roll to everyone, with metrics guiding every step.

Set dependencies between flags. Gate a feature on another flag's state. Build complex, multi-flag rollout strategies without complex code.

Release

Rollout progressively and measure the impact.

Every release is a controlled experiment. Start with a slice of traffic, watch your guardrail metrics, and expand only when the data says it's safe. If something breaks, one click kills the flag, no deployment, no incident bridge, no war room.

Ship dark, release when you're ready. Code deploys with the flag off then flip it on when you're ready. No new deployment, and no team coordination required.

Catch issues at 1%, before they reach everyone. Roll out incrementally and detect degradation the moment it appears. Flag-level attribution pinpoints the cause before 99% of users are affected.

Disable the flag in milliseconds. With metrics calculated automatically, you can confidently rollback a feature in one click without affecting the entire deployment. Guardrail metrics track performance, error rate, and conversion from the first request.

What our customers say

Teams are releasing features faster with Harness

30%

increase in mobile conversion

30x

faster releases

We've seen an increase of over 30% in conversion of our mobile usage. Harness FME has been a key part of our overall strategy ... and we're able to innovate faster because of it.

Andrew Boellstorff, Director of Digital Product and Technology, Speedway Motors

Read more
FAQ

Frequently asked questions

Harness wraps new code in feature flag statements using open source SDKs. The code is deployed to production but remains hidden behind the flag until you choose to enable it. This means you can release continuously without exposing unfinished or risky features to users.

Harness supports percentage-based rollouts, user segment targeting, and attribute-based rules. You can target by any user property your application passes to the SDK, define dependency conditions between flags, and combine rules to create precise rollout strategies.

By exposing a new feature to a small percentage of users first, you can observe real-world behavior before full rollout. If metrics degrade or errors spike, you can instantly reduce the rollout percentage or kill the flag entirely, without a code rollback or redeployment.

The Harness Feature Flag Cleanup Worker Agent integrates directly into your Harness Pipelines to automatically detect stale flags based on rollout status, targeting rules, and inactivity metrics. Teams can run scheduled scan-only audits to build trust, then trigger cleanup workflows that generate individual draft pull requests for safe human review. This pipeline-native approach ensures flag hygiene becomes part of your release workflow rather than manual tech debt sprints.

Harness includes role-based access control, approval workflows, integrated pipelines, OPA-style policies, and a full audit trail for every flag change. Teams can require sign-off before flags are enabled in production to enforce change management policies automatically.

Get started with Harness Feature Flags

Try Harness Feature Flags free. No credit card. No commit. Full access to feature flags, targeting, and dynamic configurations.