LaunchDarkly
LaunchDarkly Changelog
LaunchDarkly shifted focus toward operational resilience and observability integration. The platform now prevents event loss during authentication failures by automatically buffering SDK events to S3 with automatic replay once services recover, ensuring no telemetry gaps during outages. On the experimentation side, Guarded Releases moved to absolute difference measurements and removed manual metric thresholds, reducing false positives and stabilizing charts earlier in test runs. Flag management also gained workflow-friendly capabilities: operators can now restore previous targeting configurations directly from change history as drafts, respecting approvals and scheduled changes, while observability expanded to include Datadog Agent ingestion alongside OpenTelemetry, and user feedback notifications landed in Slack with sentiment signals and variation context.
Infrastructure and experimentation matured this month. Resilient Event Ingestion now buffers SDK events to S3 during authentication outages, preventing data loss with automatic failover and replay once services recover. Guarded Releases shifted to absolute difference measurement and eliminated manual metric thresholds, reducing false positives and improving chart stability early in experiments. Beyond core features, Datadog Agent ingestion opened observability data collection to non-OpenTelemetry customers, and User Feedback notifications graduated to Slack, letting teams triage submissions with sentiment and flag variation context directly in their messaging workspace.