cluster health: Workflows graph with 0% and green/red consistency

[tiborsimko]

Here is one Cluster Health page screenshot that looks reasonable:

Here is one that looks suspicious:

Note the workflows graph and the "red badge" there saying "0%". This is not easy to interpret: why red? (signalling stress-like error-like situations) and why 0%? (signalling total unavailability, whilst the cluster is apparently working).

Let's check how 0% is being calculated and let's improve presentation details in these kinds of situations.

• For example, if the bold text "queued" is in green, then the percentage should also be in green, when it does not signal "stress" or "error".
• Or they should be both displayed in red if it signals stress or error.
• Ditto we could colour "pending" item differently, and/or in bold, if it contributes to the overall 0% number.

If we improve colouring/bolding of text and the linking of it to the final figure, then people would be able to read and understand 0% more easily.