I have a Sensor Factory, adding the values of a specific channel of two different sensors.
When one of those sensors go down, Live Data graph shows the correct value, but the 2 days / 30 days / 365 days graphs add 10k to the amount.

Channel definition:
#1:TotalPDP
Channel(5315,0) + Channel(5366,0)

Error Handling:
Factory sensor shows warning status when one or more source sensors are in error status

If a Sensor Has No Data:
Calculate the factory channels and use zero as source value

The source sensors are SNMP Custom sensors.
Checking the graphs of the active sensor, it shows normal values (there's no sign of the 10k added to the value there).


I'm at a loss as what's causing this.
Especially the difference between the Live Data graph and the other graphs is throwing me off...
However, I need this to be fixed, since the 2 day graph is part of one of our alerting boards (and people get nervous because of the suddenly jumping graph)...

Article Comments

Hi there,

I'd like to investigate this further.

Please forward us a Support Bundle including the system log files for analysis. Please open the "PRTG Administration Tool" and send us the Log Files from the tab card "Logs and Infos". Please enter the ticket number PAE1010433 when submitting the Bundle.

Thank you!

Kind regards,
Andreas Günther, Technical Support


Mar, 2018 - Permalink

We are facing simillar issue since few days.

Live data graph show for some sensor for example value 12.926msec but in the same time on 2 days graph we see 2.767msec. How it is possible? We dont know if received Errors related with this sensor are true or false.


Jul, 2018 - Permalink

From memory, a Recalculate PRTG Graph Data Cache fixed the issue (I did it right after a PRTG upgrade round, as not to cause any additional downtime for PRTG; after all, this recalculation causes a restart of the PRTG Core Server Service)

So, our fix: - Upgrade PRTG (including complete server restart) - Recalculate PRTG Graph Data Cache (after both PRTG Cluster nodes were back up and running again)


Jul, 2018 - Permalink

Dear Sebastian,

except for the live graph, graphs in PRTG use average intervals. Short spikes will average out. For example, the 30-day graph usually shows 1-hour averages. In order to see the raw data, please use the tab "Historic Data". You can select "No Interval (Display Raw Data)" for historic data queries.


Jul, 2018 - Permalink