In a Kubernetes environment where Prometheus is being used as the primary monitoring tool, the task is to design a Grafana dashboard. The goal of this dashboard is to help identify deployments (exclude statefulsets, daemonsets, cronjobs etc) that are potentially oversized or undersized, thereby facilitating cost optimization through resource adjustments.
Oversized Deployment: If a deployment requests a large amount of CPU or memory but only uses a fraction of it, it is considered "oversized" as it inefficiently consumes resources.
Undersized Deployment: If a deployment consistently maxes out its requested CPU or memory, leading to potential performance issues, it is considered "undersized".
The aim of the dashboard is to quickly and intuitively highlight these inefficiencies, which in turn would facilitate informed decision-making and optimize resource utilization.
