Big Brother (cough cough) has a matrix that displays a group of servers and
indications of the health of such things as CPU, disk, memory, etc.

I would like to see something on the node and/or interface pages that provides a
visual cue to the health of various statistic for each device.

We could trigger on sysObjectID and calculate from datacollection-config.xml
what .rrds are available for a device. Then cross-reference this with
thresholds.xml to see what the thresholds are.

The next bit is more difficult. How do we display whether or not a threshold has
been exceeded?

The easiest way would be to drive it from events. A threshold exceeded event
turns it red, a threshold rearmed event turns it green. I know people also want
"yellow" so as we rework threshd we might keep that in mind.

A cooler way would be to not only indicate via color the condition of the value,
but to use one of Tufte's "sparkline" graphs next to it.