LaunchDarkly Accelerates Journey to Self-Healing Software with Feature-Level Observability That Reduces Developer Friction and Accelerates Triage
Session replay, APM and AI-powered diagnostics deliver observability at the point of release, giving developers tools to fix regressions faster with less friction
SAN FRANCISCO, Oct. 08, 2025 (GLOBE NEWSWIRE) -- LaunchDarkly, the comprehensive feature management platform, today unveiled key new AI capabilities and integrations that let developers ship software at even higher velocity while keeping risks at bay. These updates further the company’s vision for Self-Healing Software that allows engineers to focus on delivering amazing user experiences instead of anxiously awaiting late-night calls to fix bugs and outages.
Pressure has been mounting on developers to ship code faster and faster as organizations have moved from yearly to seasonal to agile deployments. And now, the rise of AI-assisted ‘vibe coding’ is accelerating that pace even more, increasing the volume of code shipped – and with it, the risk of shipping buggy software. Traditional Application Performance Monitoring (APM) tools, while still essential for infra and system-wide monitoring, were never built to tie the performance of the features being shipped to the underlying observability data. That has left a critical gap in observability at the release level, one that can mean costly outages and customer churn when issues go undetected.
The new observability updates, which build on the company’s previous developments from its Galaxy 2025 user conference, provide closed-loop automation for de-bugging and quality control that transform how companies approach software quality and reliability. The updates include:
- Live view of feature performance: Adding session replay, error monitoring and APM data directly into the rollout surface gives developers a live view into how changes are performing without needing to wait on legacy alerts or hunt across APM dashboards. This closes the loop between code change and customer impact, dramatically reducing mean time to repair and building confidence to ship more often.
- Regression attribution to metrics: Guarded Releases connect the dots between what changed (your feature flag) and what broke (your metric). No more guesswork, blame-pong, or digging through dashboards at 2 a.m. When a metric regresses, LaunchDarkly shows you the exact flag change that caused it and lets you roll it back with one click.
- Session replay: provides full context into how a release impacts real users right down to clicks, rage-scrolls, and form abandons. You can finally pair telemetry data with human-readable truth. It's the fastest way to diagnose what actually happens when a release behaves badly, especially when the bug doesn’t trigger an alert.
- Upcoming AI-Powered Diagnostics: The upcoming Vega AI agent will help eliminate the traditional "needle in a haystack" debugging process. Vega analyzes logs, traces, metrics, and session replays to identify root causes, generate timelines of what broke and why, and surface recommended code changes—turning noisy production data into actionable insights.
“While iterating quickly on products is paramount, organizations are becoming terrified of their own speed — they can find themselves flying blind on how their software is performing as they ship,” said Jay Khatri, Head of Observability at LaunchDarkly. “One bad release during peak season can cost millions in revenue and customer trust, which is why we’re so focused on moving from reactive damage control to proactive confidence.”
“Organizations are now building software faster than their capacity to roll it out and manage it effectively, a trend exacerbated by the rise of AI-assisted coding," said James Governor, analyst and co-founder of RedMonk. "LaunchDarkly has responded by investing in observability tooling to enable session replay, troubleshooting and rollbacks at the feature level, for stability with speed."
To learn more about the new observability features, register for the October 22 webinar featuring Jay Khatri, Head of Observability at LaunchDarkly, James Governor, co-founder of RedMonk, and JR Robinson, Senior Director of Infrastructure at Writer. To register, click here.
About LaunchDarkly
LaunchDarkly is a comprehensive feature management platform that equips software teams to proactively reduce the risk of shipping bad software and AI applications while accelerating their release velocity. By progressively rolling out features, monitoring critical metrics in real-time, instantly rolling back flawed code, easily conducting targeted experiments, and quickly iterating on AI prompts and models, development teams can ship innovation consistently and confidently. Serving over 5,500 of the most innovative enterprises, including a quarter of the Fortune 500, LaunchDarkly is trusted around the globe to deliver exceptional customer experiences and maximize business outcomes.
Media Contact: Spencer Anopol Head of PR sanopol@launchdarkly.com
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
