Datadog has unveiled Feature Flags, a feature management tool designed to enhance how engineering teams release new functionality without compromising reliability. With the launch now generally available, it integrates with Datadog's APM and RUM.
Modern feature management often operates separately from observability, requiring development teams to manually piece together telemetry to gauge the impact of feature changes on performance and reliability. This disconnection can lead to risky deployments, delayed incident responses, and increased technical debt, as unused flags accumulate across systems. Additionally, managing rollouts and rollbacks effectively can be challenging in distributed environments, often requiring custom scripts and manual oversight.
As a result, releasing new functionality can be a complex and manual process. Engineers monitor dashboards, regulate traffic with canaries or blue/green deployments, and determine whether spikes in errors or latency relate to a specific feature, configuration change, or unrelated service behaviour. This may extend the time required to launch a feature or increase the likelihood of deploying untested features.
Feature Flags by Datadog aims to address these challenges by linking each feature flag to real-time observability data. Through this integration, teams can identify reliability issues associated with a specific feature or configuration. The tool seeks to support automated rollouts and rollbacks, enforces experimentation guardrails, and helps clean up unused flags before they contribute to technical debt. It complements Datadog’s CI/CD and test optimisation products by extending observability into release management.
"Releasing new features is one of the riskiest parts of modern software delivery, and releasing frequently is even more important in today’s AI-driven development age,” said Yanbing Li, Chief Product Officer at Datadog. “Datadog Feature Flags, created with a head start after our acquisition of Eppo, allows development teams to automatically detect regressions, enforce reliability guardrails, and ship updates faster and more safely by tying every flag to real-time telemetry.”
Datadog Feature Flags aims to help organisations deliver new functionality by providing:
- Unified Observability and Feature Management: Correlate each feature flag with Datadog telemetry to see its impact on performance and reliability.
- Automated Rollouts and Rollbacks: Mitigate risk with canary releases, circuit breakers, and instant rollbacks informed by real-time service health signals without manual intervention.
- Dynamic Configuration and Safe Experimentation: Adjust system behaviour instantly without redeploying code, maintain guardrails, and support reliable experimentation.
- Automated Stale-Flag Cleanup: Reduce technical debt with integrations that identify unused flags and initiate safe removal of inactive paths from codebases.
