Sift is a machine-learning-powered diagnostic feature in Grafana Cloud designed to assist SREs and DevOps teams in automating various aspects of incident investigation, such as detecting new errors, identifying recent deployments, and recognizing overloaded Kubernetes nodes. Since its public preview launch, Sift has introduced several new features, including an improved homepage, expanded investigation capabilities beyond Kubernetes, and a new outgoing webhook in Grafana OnCall for automatic investigations. Users can now trigger Sift investigations directly from various Grafana pages, with options to configure checks and customize settings for more effective diagnostics. A notable addition is the HTTP Error Series check, which identifies increased HTTP error rates and provides visual metrics for quick problem identification. Sift's investigation timeline feature correlates events for easier analysis, and the platform continues to evolve with plans for enhanced label selection and new diagnostic checks. Available at no extra cost in all Grafana Cloud tiers, Sift aims to streamline incident response and reduce mean time to recovery (MTTR).