ADS-B tracks spike whenever I change gain and restart dump1090-fa. I reduced gain from 49.6 to 46.6. Percent of strong messages went from 3.3% to 2.3%. Is this hardware related or simply a statistical quirk?
Statistical quirk; when you restart dump1090, it does not preserve information about currently tracked aircraft. So all the currently-visible aircraft are counted as “new” after the restart, producing the spike.
I know there’s an option within the graphs configuration to set a maximum value which helps but I removed that when we had the tropo opening last week because that graph was pegged fully green.
This can be frustrating though.
I wonder of there’s any way for the graphs to do a ‘sanity check’ on the data and if it suddenly shots up massively for a few seconds and then drops down, then that spike is ignored?
I’m speculating, this may not be possible/practical.
Once hacked my dump1090-fa not to emit any stats.json for the first minute.
That fixes it as the first number after the counter decreases doesn’t count.
I suppose you could stop collectd, restart dump1090-fa, wait 2 minutes and then restart collectd.
Since about 10/30/2019 graphs1090 has not been showing any 978UAT traffic. 978UAT shows up in tar1090. What can I do to remedy this? Thanks!
Probably the switch from skyview to skyaware naming.
Have you tried reinstalling graphs1090?
Hello friends. Since last week I noticed that my Graphs1090 Message Rate graph shows Min/Max spikes (light-yellow curve) around 23:45 (see image below). It’s strange because the spikes of previous days disappear when I select 48h, 7 days, 48 days, and so on scale. Some idea?
The yellow shaded area bounds the minimum and maximum message rates seen over the preceding 7 days (ie not including the current day). At some point during the past 7 days you had some anomaly that caused a brief spike - perhaps a restart of dump1090 or something like that. It will disappear once that event is longer than 8 days ago.
That shaded area only covers a day, so on longer period graphs it will only show at the right side, not the whole graph. That particular spike may not appear on longer period graphs because of the way the data is stored in collectd. It’s a round robin database which has different resolutions for different time periods in order to reduce the space taken up. The highest resolution data is only present for a few days (which has a data point once per minute I think). As the data ages it gets reduced in resolution meaning that short duration events get averaged out and disappear from the graph.