Hi, we had the problem that our core server was poorly accessible for about 5 hours due to network problems. So all remote probes were not able to transfer data correctly or at all to the core server. Now we have a gap on the core server over this 5 hours on all sensors. I thought that the remote probes are caching it's data for that 5 hours and submit the cached data when the core server is reachable again. One remote probe has 5 Sensors with low frequented scanning interval (30 seconds for NetFlow V9, SNMP Traffic and Core/Probe Health, 30 minutes for 2 windows services), so there's not much data to cache I guess. Currently we have 70 Remote Probes online.

Any Ideas why the missing data in the mentioned 5 hours was not cached by the remote probes?

greetings, mike


Article Comments

Dear Mike,

PRTG's Remote Probes can buffer a maximum of 500,000 sensor results in RAM memory of the remote probe system (this is up to 50 - 200 MB of monitoring data). However, data is only stored in RAM memory. In case the system running the Remote Probe is rebooted, monitoring data in the buffer gets lost. I assume this was the case here.


May, 2011 - Permalink

No because none of the 70 remote probes have submitted any results. It's not likely that all servers did a reboot within that time - I'm actually sure. In my opinion either caching didn't work or sending/receiving the cached data was unsuccessful. Is there something to configure for caching the results?


May, 2011 - Permalink

This might be just a caching issue. Please try the following:

  • Stop both PRTG services (core and probe).
  • Head to your data folder (can be discerned from the PRTG Server Administrator tool, under the Core Server tab)
  • Delete the PRTG Graph Data Cache.dat file
  • Restart the PRTG services
  • Log in to the Web GUI
  • Go to Setup | System Status and check the background tasks. It should state that it is calculating historic data

Then,

  • Refresh the page from time to time (is the number being reduced?)
  • Let it run until it is done and then take a look at the historic data of your sensors - it should now be available

May, 2011 - Permalink