CloudBees Acquires Rollout: Adding Feature Flag Management to Reduce Deployment Risk

Originally announced June 11, 2019. Source: devops.com / CloudBees press release.

CloudBees Brings Feature Flag Management Into Its CI/CD Platform

CloudBees announced on June 11, 2019, that it had acquired Rollout, a feature management company whose platform gives developers and product teams fine-grained control over when specific features become visible to end users. The acquisition extended CloudBees’ continuous integration and continuous delivery (CI/CD) portfolio by addressing one of the last high-risk steps in software delivery: the moment a new feature is actually exposed to production traffic.

Why Feature Flags Matter in a CI/CD Pipeline

Continuous delivery solves the problem of getting code into production quickly, but it does not on its own control when users experience a new capability. Feature flags — sometimes called feature toggles — solve that gap. A flag is a conditional in the codebase that routes traffic to either the old or new code path based on rules defined at runtime, without requiring a redeployment.

This separation of code deployment from feature release has several practical advantages for engineering and product teams:

• Incremental rollouts: Expose a new feature to 1%, 10%, or 50% of users before a full release, limiting blast radius if something goes wrong.
• Instant rollback: Turn a feature off at the flag level without reverting a deployment or executing a hotfix release cycle.
• Production testing: Run A/B experiments or canary tests against real traffic to gather behavioral data before committing to a feature.
• Trunk-based development: Developers merge incomplete features behind a flag, keeping the main branch always releasable.
• Cross-platform control: Flags can govern feature exposure on web, backend services, and mobile clients from a single control plane.

Forrester Research noted in a January 2019 report (AD&D Pros: Stop Guessing And Start Experimenting, Christopher Condo and Bill Seguin) that “getting ideas in front of users faster — ‘continuous discovery’ — is the new normal,” and that feature flags give application development and delivery teams “complete control of when to expose a particular feature to users, allowing administrators to turn it on and off at the flip of a switch.”

What Rollout Brought to the Table

Rollout positioned itself as a developer-first feature management platform with an emphasis on security and ease of integration. Key capabilities included:

• SDK-based feature flag implementation compatible with multiple languages and mobile platforms.
• Targeting rules that let teams segment rollouts by user attributes, geography, account tier, or custom properties.
• Audit trails and governance controls suited to regulated or enterprise environments.
• Instant kill-switch rollback that operates independently of the deployment pipeline.

These capabilities complemented CloudBees’ existing Jenkins-based CI/CD tooling by closing the loop between pipeline automation and controlled feature exposure.

Strategic Rationale for CloudBees

Sacha Labourey, CEO and co-founder of CloudBees, framed the acquisition around risk reduction: “Our goal is to help organizations deliver great, feature-rich software efficiently while minimizing risks associated with the deployment process. The acquisition of Rollout gives CloudBees customers the flexibility to decouple features from software versions.”

Erez Rusovsky, CEO and co-founder of Rollout, joined CloudBees as product director for the Rollout offering. He described the combination as additive: “Feature flags and CI/CD combined help companies truly become continuous and increase the pace of innovation.”

From a market positioning perspective, the deal signaled that end-to-end software delivery management must extend beyond pipeline automation to include runtime feature control. Organizations practicing trunk-based development or pursuing high deployment frequencies need a governance layer that keeps releases safe after the pipeline finishes.

How Feature Flags Fit Into a Modern DevOps Workflow

A typical integration of feature flags into a CI/CD pipeline looks like this:

1. Development: The engineer wraps new functionality in a feature flag. The flag defaults to off, so the code ships without user impact.
2. CI: Automated tests run against both flag states (on and off) to prevent regressions in either code path.
3. CD / Deployment: The artifact is deployed to production with the flag still off. No user-visible change occurs.
4. Controlled release: The product or ops team enables the flag for an internal test group, then expands to a percentage of users, monitoring error rates and business metrics.
5. Full release or rollback: If metrics are healthy, the flag rolls out to 100% and is eventually removed from code. If something breaks, the flag is turned off instantly—no deployment required.

This workflow eliminates the “big-bang” release model, where all users receive an untested change simultaneously, in favor of a controlled drip that surfaces problems early with limited exposure.

Context: CloudBees’ Broader Platform at the Time

At the time of the acquisition, CloudBees described itself as the largest contributor to Jenkins and Jenkins X and a founding member of the Continuous Delivery Foundation (CDF). Its customer base included over 3,500 organizations and more than 50% of the Fortune 500, according to the company’s own figures. The Rollout acquisition was part of a broader strategy to build what CloudBees called a “software delivery management” (SDM) system that covers the full lifecycle from code commit to production feature exposure.

Key Takeaways for DevOps Teams

• Feature flags are a risk management tool as much as a developer productivity tool; they decouple the binary choice of “deploy or don’t deploy.”
• Integrating flag management into a CI/CD platform (rather than bolting it on afterward) enables consistent governance, audit logging, and policy enforcement across the delivery pipeline.
• The acquisition reflected growing industry recognition that deployment frequency alone is not sufficient—controlled feature exposure is the final mile of safe continuous delivery.
• Teams evaluating feature flag platforms should assess SDK coverage for their target platforms (web, mobile, backend), targeting granularity, rollback latency, and integration depth with their existing CI/CD tooling.