cilium/hubble ? reverse-engineered prompt

Build me a Kubernetes observability tool called Hubble that gives me a clear view of service traffic, network health, and security events.

I want it to let me see which services are talking to each other, what DNS and HTTP activity is happening, where requests are failing, and when network policies are blocking traffic. It should feel like a CLI first project, with flow inspection commands, useful summaries, and the ability to filter by namespace, pod, time range, protocol, and status. If it makes sense, include support for viewing metrics and a simple service map style output too.

Make it something that can run with Cilium and use eBPF based visibility, and keep the setup practical for a Kubernetes cluster. If you need to check current docs or best practices for the latest Hubble and Cilium setup, look them up online. Include the basic install and getting started flow so someone can try it quickly.