I've got three SNMP Traffic 64bit sensors that all say something like this - "1 hour interval average of 47 kbit/s (Traffic In) is unusually low for this hour of the week"

All three say this in "Last Message". However, they have all said the exact same unchanging message for the last three days. The "interval average" they report is not current.

So I want to know, what time did this happen? When was "this hour of the week"? How can I clear this message?

Article Comments

Interesting... I just changed the system Graph Settings, which required a reboot of the PRTG server. After this reboot, the Unusual items mentioned above went away.


Apr, 2014 - Permalink

Hello,

thank you very much for your KB-Post. The Unusual Detection always compares back to the same time of day (and day of the week), into the past. And for this comparison it then uses an hourly average (of this time of day, day of the week). Against this average, the last value is compared, and if the value differs too much, the unusual state is triggered.
You can change the "thresholds" in the System Settings. If the current values show a difference above the threshold, the sensor will probably show the unusual state for at least this one hour (as the base is a hourly average), until either the current values go "back" to normal, or there is another average.
Changing the Graph Settings actually causes a recalculate of the "Graph Cache", during this process the Unusual Detection will be paused.

best regards. best regards.


Apr, 2014 - Permalink

Okay, well, something obviously got stuck, because the message remained for three days and the value in the message did not change during those three days even though the traffic varied widely. That is why I was very confused - it kept saying "this hour", but that clearly wasn't true.

This must have been a stuck message leftover from a bug because it cleared immediately upon reset of the PRTG server.


Apr, 2014 - Permalink