When pods scheduled to a node have limits, the node dashboard is confusing in how it conveys those limits.

The attached screenshot shows a master with 2 pods that request resources for memory and CPU. The text is

3.05 available of 150m total limit - Usage 944m

That does not make sense. The node:

• has 4 cores (capacity)
• has 3500m (3.5 cores) (allocatable)
• has 150m of pods requesting limits (which is almost irrelevant here)
• has 944m core of usage
• has an 1.79 requested cpu (in graph)

I expected to see:

<total_available: capacity - usage> of capacity, <total_pod_request: allocatable - usage> requested

with the graph showing lines at 4 cores (total capacity), 3.5 cores (allocatable), 1.79 (sum of pod request), and usage (944m), and probably a limit line (of 150m) which is a wierd number anyway but it's fine.

I also expected to see one graph entry on the node dashboard for all allocatable resources:

allocatable:
attachable-volumes-gce-pd: '127'
cpu: 3500m
ephemeral-storage: '123201474766'
hugepages-1Gi: '0'
hugepages-2Mi: '0'
memory: 14226848Ki
pods: '250'

and the graph showing the requests over time for any we don't recognize (go through cpu, filesystem, memory, network, then show the rest of the counts against their limits). Anything '0' can be ignored. For instance, i expected to see attachable-volumes-gce-pd show up.

In addition, clicking on the usage takes way longer than I expect (also wierd, because the query should be able to return nodes directly - nothing needs ot be joined).

It's worth noting that nodes != overall - the set of things in the overview dashboard we'd show don't match what nodes should show.