Show the feed that stopped reporting

The chart that worries me most on a monitoring screen is the empty one. A feed that has gone quiet looks, at a glance, exactly like a system with no incidents. Within a week, operators stop seeing it at all.

In SysHealth that silence had to count as a finding. AIX, Linux, storage, and network sources arrive in different shapes, and some come from a command rather than an exporter, so a quiet feed can mean anything from “all fine” to “the job stopped running”. The screen can't treat those as the same thing.

The front page I aim for is short: what is reporting, what has gone quiet, and what is still open. A quiet feed should look different from a healthy one, and it should say when it was last heard from. Raw output stays one step away for whoever is digging in.

I'd rather show a shorter list that admits a hole than a green wall that hides one. That's about how the product should behave, not a claim about any estate's uptime.