Actually, the standard PRTG 6 graphs were identical to the standard PRTG 8 graphs. Both graphs average values over time. In PRTG 6 it was possible to generate unaveraged graphs using the View -> Historic Data function. This option can now be found in PRTG 8 under Sensors -> View Historic Data. Beyond that, it is possible to define shorter scanning intervals (down to 1 second) under PRTG 8 (Setup -> System Administration -> System & Website) under the "Scanning Intervals" options. It is, however, not recommended to use such small intervals, particularly when monitoring a large number of sensors, seeing as this can cause issues not only with the monitoring functionality but also result in bandwidth saturation (each scan uses network resources, so the more requests are forwarded within a specific period of time, the more these resources will be necessary for the software's usage).
Last but not least: although small scanning intervals result in less "smoothed" graphs, even when using a small scanning interval the same are still being averaged. Only the historic data option (using raw values) displays the actual monitoring points as received by PRTG. How this affects the graphs of course also depends on the sensors used. Small scanning intervals might make sense with SNMP sensors, for example, but would certainly be overkill when monitoring NetFlow or Packet Sniffer sensors.
We personally recommend keeping your scanning intervals above 10 seconds. I'm afraid we can't offer support for smaller scanning intervals, seeing as these can lead to issues, as mentioned above. Nonetheless, you can certainly define and use these intervals.
Add comment