On startup (and during errors?) the graph freezes and fails to update even though pings continue. Same graphic now for the past two boots (cold, not hibernate)

I started writing a feature to toggle the display of a running average of the pings in the current log. However, it wasn't immediately clear to me how error pings should be handled.

Simply ignoring them doesn't feel accurate, as it would make the average higher than it should be, and probably defeat the purpose of displaying it. (Keeping an eye on your network quality)

But how do you numerically account for the missing ping in an average?

Perhaps a running average feature just doesn't make sense in this context.

https://github.com/jaredks/rumps

This was written for Unity, needs to be modified for the modern Ubuntu tray

• still contains debugging timing
• "start automatically" label not updated properly

I'm thinking something like Cairo for drawing 2D graphs ought to work. There's even plotting libraries made from it, with python bindings.

http://cairoplot.sourceforge.net/index.html

For instance, Indicator Multiload uses Cairo to draw the performance graphs:

https://launchpad.net/indicator-multiload