Stack monitoring: Unit mismatch in memory graphs

[afgomez]

The memory graphs shown in the stack monitoring plugin have a mismatch between the unit specified in the header and the values shown in the graph.

Example in APM server view:

Example in Kibana instance view:

As a user I would expect the scale of the numbers in the graph to match the unit in the header. Taking the screenshots as an example, I would expect the APM server graph to show numbers in the range of 120, 100, 80... instead of millions.

[ravikesarwani]

The unit in the header says "MB" or "GB" but the Y axis plots "bytes". Shouldn't it convert and plot as MB/GB? Looks like we are already doing all the work to define the correct unit based on min and max values in the range.
Once we have finished 7.9 tasks we should take a look to see how we can potentially fix this. Seems like a bug to me.