You are working on a large Kubernetes deployment with hundreds of nodes and thousands of pods. You are using Prometheus for monitoring, but as the cluster grows, you are starting to notice that Prometheus is consuming a lot of memory and CPU, and the response times for queries are becoming slower. You need to ensure that Prometheus scales effectively with the cluster while maintaining the performance of your monitoring solution. Which of the following is the most appropriate solution to scale Prometheus in a large Kubernetes cluster?