Tee82-debug/intelligent-power-monitoring-system ? reverse-engineered prompt

Build me a real time power monitoring and infrastructure observability prototype for a small Kubernetes cluster.

I want it to collect electrical readings from a smart plug on the main node, collect node and pod metrics from k3s, store everything in Elasticsearch, and show it in dashboards so I can see power use, CPU, memory, restarts, and overall health in one place. It should also detect unusual power behavior, make short term power forecasts, and keep separate views for raw data, anomalies, and predictions.

If it makes sense, include a small FastAPI service for AI assisted cluster questions and a simple health prediction endpoint based on the collected metrics. Make the app feel like something an operator could actually use, with clear status, alert handling, and easy to understand charts. Use current docs online if you need to, and set it up so it can run locally in a Docker based dev environment.