I encountered an issue where incorrect metrics were being logged in Prometheus within a Kubernetes cluster. These metrics were exposed from a Python-based REST API service which was running in a multi-pod environment built using the FastAPI framework...