2024 has been a year of hard lessons for teams that build software—poorly executed rollouts, failed app launches, and large-scale outages that have not only disrupted businesses but also affected millions of users. It’s also been an exciting year of innovation in technology (hello, Generative AI!), but as we’ve seen in some painful examples: moving fast often breaks things.
LaunchDarkly believes the way most teams release software is just plain broken — a rush to ship features and nothing to do but hope for the best. And this challenge is only magnified when it comes to building and shipping AI applications or trying to do something truly novel for your customers. But what if there was a way to move fast without sacrificing quality or performance? And what if–even better–this could all happen automatically?
This world is possible with Guarded Releases. Read on to learn about our brand new capabilities to guard your releases, automatically remediate incidents, manage risk across your release process and team, and build AI applications with more control.
Introducing Guarded Releases
A staggering 80% of software bugs are caused by new code changes—risks that every development team faces when pushing updates. We’re here to make risky code changes a thing of the past.
LaunchDarkly is setting a new standard in software delivery with Guarded Releases, ensuring that every release is monitored with the ability to quickly roll back at the first sign of trouble and allowing teams to innovate quickly without compromising reliability or the user experience:
- Performance safeguards and automatic rollbacks: Set thresholds to catch issues early, triggering instant rollbacks to a stable state when limits are breached.
- Phased rollouts with guarded progressive releases: Gradually release updates in stages, with the ability to pause or roll back if issues emerge.
- Insights into release health: Get end-to-end visibility into the health of your releases with real-time insights into regressions and their resolution status.
- Real-time alerts with PagerDuty notifications: Get rapid, actionable alerts on regressions and recovery status to stay on top of release health.
- Enhanced usability and visual insights: An updated interface with clear metrics, streamlined rollbacks, and hovercard explanations supports quick, data-driven decisions.
- Plus, error monitoring and session replay (coming soon): Instant error tracking and alerts for faster fixes and improved reliability. Replay user actions to understand where they experienced issues.
Want to learn more? Check out our in-depth blog post.
Build better (and safer) AI apps
The rapid evolution of LLMs and GenAI has opened up a world of possibilities for software teams across industries. From GPTs that have fundamentally changed the way we write, to AI assistants that help us manage everything from calendars to software production, the applications are vast and transformative. But bringing non-deterministic AI applications to production at enterprise scale presents unique challenges, and tooling has yet to adapt in many cases. Currently, AI engineers have limited options, and often have to manually update configs and model parameters in code, which requires frequent redeployment and slows down the development cycle. That adds a lot of friction to the process of finding the optimal model configuration, paired with the right prompts to generate a desired outcome—repeatedly, and at scale.
To power the next wave of AI builders and products, we're excited to announce AI configs, now available in Early Access. AI configs are purpose-built to manage the configuration of your prompts, models, and configurations that power your GenAI application, which can be updates, upgraded, or experimented on at runtime. You can use them to:
- Create and manage prompt and model configs (AI configs) at runtime
- Instantly roll back to a stable configuration in case of issues
- Visualize and monitor for regressions on key metrics like token usage for confidence
- Target AI Configs to various audiences to customize the end-user experience
- Run multiple configurations against the same application to track performance changes
Want to learn more? Check out our in-depth blog post.
New ways to manage releases at scale
It’s well-established that feature management is a best practice for high-velocity engineering teams to reduce risk, improve productivity, and enhance customer experiences. But like any practice, with scale comes additional complexity.
We’ve heard from customers that scaling feature management leads to questions like: How can I more efficiently clean up stale flags? How do I know if my team is following feature management best practices? And how can I integrate LaunchDarkly into my existing change management platform?
That’s why we’re shipping several new features that make it even easier to efficiently manage releases across your organization:
Launch Insights Dashboard
The Launch Insights dashboard, available now for users on every plan, is a project-level view that gives engineering leaders a better understanding of how well your team is managing software release risk, embracing best practices, and optimizing customer experience. The dashboard shows your team’s usage across the LaunchDarkly platform, which is grouped into the categories of Release, Monitor, and Optimize. You’ll see an overall score, as well as a score and summary for each specific category.
The dashboard also shows trend information like how many flags are ready to archive, so you can understand how well your team is managing flag cleanup. Plus, it includes information on the top flag creators across an organization, which you can use to share learnings from top users.
As your teams adopt processes like progressive releases, guarded rollouts, and experimentation, you’ll see your scores go up, and can use the dashboard to give clear guidance to your teams on ways to improve their software release processes.
Flag cleanup made easy
To help solve the problem of tech debt from stale flags, we’re introducing automated checks in the Flags list to determine if a flag is ready to archive or ready for code removal. That means that your team can more easily and intuitively manage flag debt, contributing to a cleaner codebase and freeing up more time for building and shipping new products.
Custom Role Builder
Using the new Custom Role Builder, LaunchDarkly users can more easily create custom roles and policies for custom roles. Coming soon, we’ll introduce Global Roles as a way to reduce the number of custom roles your organization has, create roles more quickly, and more easily manage multiple teams across large organizations.
Updated ServiceNow integration and custom approvals integrations
With our new ServiceNow integration (available to Enterprise plan customers), you can automatically approve and apply changes in LaunchDarkly when they’ve been reflected in ServiceNow. If you have complex workflow management processes or approval requirements, you can also create your own custom approvals integration in LaunchDarkly.
Want to learn more about these new features? Check out our in-depth blog post.
Ship what's next
We hope our latest release accelerates your team in building your next great product. Ready to explore LaunchDarkly or see our newest product updates up close? Schedule a demo today.
.webp)
