One problem I am having is that graphs, especially ones with filled areas, tend to be misleading in periods of no-data.
For example, I've set up ping response data (loss and latency) to not record data when the node is unreachable. But if I plot those, during a period with no data recorded the lines just go straight across between the last two good point and the the first when data is restored.
Are there any configurable (even if needed to be compiled in or patched) controls over this?
In particular, two possible options: Lines and filled regions that included a data point as well, to indicate the actual data collected, or just omitting the line/region when there is more of a gap than X (I realize defining x may be hard in general).
Any ideas how one might address reflecting no-data for multi-item graphs in this regard?
I haven't looked at the code yet, is this all from some general package where there may be additional options one could turn on in code easily (like putting points on the lines/region-boundaries?)
For example, I've set up ping response data (loss and latency) to not record data when the node is unreachable. But if I plot those, during a period with no data recorded the lines just go straight across between the last two good point and the the first when data is restored.
Are there any configurable (even if needed to be compiled in or patched) controls over this?
In particular, two possible options: Lines and filled regions that included a data point as well, to indicate the actual data collected, or just omitting the line/region when there is more of a gap than X (I realize defining x may be hard in general).
Any ideas how one might address reflecting no-data for multi-item graphs in this regard?
I haven't looked at the code yet, is this all from some general package where there may be additional options one could turn on in code easily (like putting points on the lines/region-boundaries?)