DEVOPSdigest

LaunchDarkly announced a new approach to software delivery—Guarded Releases—that empowers organizations to ship with confidence and manage risk proactively.

At its core, a Guarded Release strategy incorporates essential elements that allow engineering teams to move quickly, including on AI projects, without breaking anything. By prioritizing application thresholds, real-time monitoring and proactive issue resolution, teams can detect and address problems before they escalate into outages. This strategy also enables business leaders to take ownership of code quality by appending meaningful metrics to every release.

“Guarded Releases mark a huge leap forward in how we deliver software,” said Dan Rogers, CEO of LaunchDarkly. “For too long, the industry has operated under the belief that speed necessitates risk. This mindset is no longer acceptable. By implementing Guarded Releases, teams can innovate confidently, and business leaders can rest easy, assured that their code quality is secure.”

Key Elements of a Guarded Release:

- Application Performance Thresholds: Establishing benchmarks for performance, latency, and availability at every rollout stage.

- Release Auto-Remediation: Responding to errors within milliseconds, rolling back to the last stable state automatically when performance thresholds are breached, minimizing disruption.

- Release Monitoring: Gaining end-to-end visibility into the health and risk of releases, with real-time insights into regressions and their resolution status.

During its Launch Week, LaunchDarkly introduced Guardian Edition, its most advanced offering yet. Guardian Edition brings the three core elements of a Guarded Release—Application Performance Thresholds, Release Auto-Remediation, and Release Monitoring—along with Error Monitoring and Session Replay, making it easier for teams to adopt Guarded Releases. With Guardian Edition, companies and their teams can help ensure that every software release is safe, monitored, and secure.

“Organizations are struggling with the increasing complexity of software builds and deployments, and constant iteration, with a lack of insight into the software release process.” said James Governor, principal analyst and co-founder of Redmonk. “Guardian Edition is designed to fill that gap, with integrated real-time insights to close the feedback loop, automating the end-to-end process of Progressive Delivery.”

Also during Launch Week, the company introduced AI Configs, enabling engineering teams to manage LLM models and prompts at runtime, target configurations to different audiences, and monitor for key metrics like token usage. This gives teams building AI software the ability to quickly and safely iterate on prompt and model configurations and customize user experiences while tracking performance.

Additionally, LaunchDarkly unveiled automated flag archival checks to manage feature flags no longer in use and the associated tech debt, simplifying the archiving or removal process for cleaner codebases. With the Launch Insights dashboard, users can quickly view adoption metrics for risk-free release practices, tracking release targeting, metrics monitoring, and experiment usage. An updated integration with ServiceNow better enables enterprises to use it as their system of record, enhancing scalability and governance across releases.