Smarter multi-cluster scheduling with dynamic scoring framework

The Dynamic Scoring Framework is an open source project that automates cluster scoring for multi-cluster Kubernetes environments using Prometheus metrics. It bridges monitoring systems and the Open Cluster Management (OCM) Placement API through three components: DynamicScorer CRD (defines scoring logic), DynamicScoringConfig CRD (aggregates and distributes configs), and DynamicScoringAgent (deployed per cluster to fetch metrics and publish scores). Developers write custom scoring APIs—for cost, latency, or predictive CPU load—while the framework handles metric collection, API calls, and PlacementScore updates. Tools like Kueue and ArgoCD that already integrate with OCM's Placement API benefit automatically. A step-by-step setup guide covers Helm installation, Prometheus setup, deploying a sample scorer, and verifying AddOnPlacementScores.