Optimal service scheduling in Kubernetes is crucial for reducing operational costs, improving resource utilization, and freeing up compute resources for more customers. It becomes a complex problem in the case of applications with services that exchange high traffic rates. High-affinity services must be grouped and placed in the smallest possible number of Virtual Machines to maximize resource utilization and minimize Egress traffic. In this work, service placement is formulated as a graph clustering problem. Score is a Kubernetes platform of state-of-the-art graph theoretic algorithms, service mesh, and metric collection tools that facilitate service scheduling. The Maximum Standard Deviation Reduction (MSDR) clustering method demonstrated enhanced long-term optimization capabilities, albeit with increased runtime complexity.