Datadog adds feature flags to strengthen software reliability

Datadog, an AI-powered observability and security platform for cloud applications has launched Feature Flags, a new capability that integrates feature management with observability to help engineering teams release software updates more quickly while maintaining system reliability. The product is now generally available and is integrated across Datadog’s application performance monitoring (APM) and real user monitoring (RUM) tools.

In many organisations, feature management operates separately from observability systems. As a result, development teams often have to manually connect performance data with feature changes to understand their impact. This lack of integration can lead to blind spots, risky deployments, slower incident response, and the gradual build-up of technical debt as unused feature flags remain in production environments. Teams may also struggle to coordinate rollouts and rollbacks across complex systems, relying on custom scripts and manual processes to manage risk.

These challenges have made the release of new functionality a high-pressure and time-consuming task. Engineers typically monitor dashboards closely, introduce changes gradually through canary or blue-green deployments, and work to determine whether increases in errors or latency are linked to specific features, configuration updates, or unrelated system behaviour. As a result, launching new features can take weeks or months, or may proceed with higher levels of risk.

Datadog’s Feature Flags is designed to address these issues by linking each feature flag directly to real-time observability data. This connection allows teams to trace performance or reliability problems to the precise feature or configuration responsible. It also supports automated rollouts and rollbacks, helps establish guardrails for experimentation, and enables the removal of outdated flags before they contribute to long-term maintenance challenges. The new capability extends Datadog’s existing CI/CD visibility and testing tools into the release management process.

Yanbing Li, Chief Product Officer at Datadog, said that releasing new features remains one of the most challenging aspects of modern software delivery, particularly as development cycles accelerate. He noted that Feature Flags, developed following the company’s acquisition of Eppo, is intended to help teams detect regressions, apply reliability safeguards, and link feature releases more closely to real-time system data.

With the introduction of Feature Flags, Datadog aims to provide engineering teams with a more integrated approach to managing software releases. By combining observability with feature control, the company is seeking to reduce deployment risks, improve operational visibility, and support more predictable and controlled innovation in complex, distributed environments.