Unleash 7 Webinar Recap: In Case You Missed It

Last week, we celebrated our latest major release with a webinar. For a deeper dive into all the demos and discussions, you can watch the full recording on our YouTube channel. In the meantime, here’s a look at what we covered.

Introducing FeatureOps

While DevOps has revolutionized how we ship software, making it more reliable and efficient, it doesn’t fully address the complexities of coordinating and validating what we ship. This is where FeatureOps comes in. It’s a set of practices that gives you runtime control over your application, optimizing for quick feedback loops and data-driven decisions.

FeatureOps is built on four key pillars:

1. Controlled feature release: Use feature flags to control who sees new functionality and when.
2. Surgical rollback: Instantly disable only the faulty part of a release without rolling back everything using a code deployment.
3. Full-stack experimentation: Validate improvements by measuring their impact on the business, engineering, and customer metrics.
4. Lifecycle management: Establish processes for managing flags at scale, from creation to cleanup.

From A/B Testing to Full-Stack Experimentation

Most of our daily work involves improvements where the “what” is clear, but the “how well” is not. What better looks like depends on what data source we look at. Full-stack experimentation addresses this by evaluating changes across three critical dimensions:

• The Voice of the Business: Are we hitting our KPIs like conversion or adoption?
• The Voice of Engineering: Is the new feature introducing errors or increasing latency?
• The Voice of the Customer: Are we solving a real problem in a meaningful way?

Solutions Engineer Nick demonstrated how Unleash connects to your existing toolstack (like Grafana, Mixpanel, and Sentry) to gather data from all three “voices.” By running an experiment with a new chatbot feature, he showed how to correlate a poor user experience with increased engineering costs and negative business impact, all without moving data to a new silo.

Looking ahead, Ivar introduced Impact Metrics, an upcoming feature that will bring key metrics from various sources directly into Unleash, creating a holistic overview of a feature’s success and paving the way for automated rollouts.

Managing Flags at Scale

As organizations adopt feature flags at scale, keeping everything organized becomes a major challenge. Our new Feature Lifecycle Management capabilities provide a structured process to guide flags from creation to cleanup.

Our UX Designer, Anniken, showcased how to streamline rollouts with Release Templates. You can now define a series of rollout milestones (e.g., “Internal Users,” “20% of Customers,” “100% Rollout”) and apply this template to any feature flag with a single click. This helps create a consistent and streamlined process across all your teams and feature rollouts.

To help you remember the crucial cleanup step, Unleash now provides helpful reminders, suggesting when a flag might be ready for completion or archiving. We’re also exploring integrations to automate cleanup by generating Jira tickets or GitHub pull requests.

Adding to the cleanup toolkit, we had a quick demo of External Links. You can now link feature flags directly to resources in other systems, like a Jira ticket, a design in Figma, or a GitHub code search showing all usages of that flag. This ensures that feature flags are always tied to the work that created them and helps with organization and cleanup.

Performance and Resilience for Global Enterprises

For global companies, performance and resilience are non-negotiable. The feature flagging system cannot be a single point of failure. That’s why we built Unleash Enterprise Edge, a high-performance, read replica of your Unleash server that scales to millions of applications and ensures your flags are always available.

Key features include:

• Hosted by us: We’ll manage Edge for you in any of 11 AWS regions worldwide for a high-resilience, global network out of the box.
• Streaming enabled: Get near-real-time feature flag updates to your SDKs without the need for polling.
• Observability: View real-time performance and health metrics for your Edge instances directly within the Unleash UI.

Taming AI: Using Unleash to Manage AI-Generated Code 

AI is changing how we write code, but with increased adoption comes a decrease in delivery stability. More bugs are making their way into production faster. Feature flags are a critical control mechanism to manage this risk.

We outlined 11 ways to leverage Unleash with AI, including:

• Testing different AI models using variants.
• Gradually rolling out AI-powered features to specific user segments.
• Using kill switches to instantly disable misbehaving AI.
• Applying role-based access control and approval workflows to AI agents.
• Using full-stack experimentation to measure the true impact of AI-generated code.

When things inevitably go wrong, you need to act fast. Platform Lead, Saul, demonstrated the new Event Timeline, a dashboard that gives you a chronological view of all changes made across your Unleash projects. In an incident, you can quickly see which flag was changed right before the issue started, identify the owner, and disable it instantly to resolve the problem.

This latest release is packed with features designed to help you ship faster and smarter, from managing technical debt to validating the impact of every change.

If you want to hear more best practices of using feature flags at scale from leading tech teams, register for UnleashCon, our annual user conference.

To stay up-to-date on new features, future events, and everything else happening in the world of Unleash, sign up for our newsletter.