feat: smokeping like latency graphs

[mfreeman451]

Since for every point we actually have a tuple of information (3 pings) we could consider keeping this in the database and using it to generate a stacked area chart for nodes. I don't think it would make sense to try and triple the size of our ring buffer and potentially kill our implementation to have it in memory since we're really only using that for the spark graph.

We should probably do some work to see how much of an increase in the payload size we would be dealing with and what kind of performance impact that would be.