I looked through a few articles, but did not get a real good feel for the Internet bandwidth requirements.

We are running a single server at out CoLo. It is 90% front end server for 20 remote probes. It has a few ping sensors for monitoring the outside connection from the CoLo.

I have calculated that our hardware can handle up to 50,000 sensors, but what kind of web server performance will I see? I changed the live data to 1 hour already and we are running 5,000 sensors. Is there a good formula for Remote probes * sensors = throughput required?

Article Comments

Hello,

for the bandwidth-"needs" of a remote probe please see How much bandwidth is required for remote probe and core server connections?.

I'm afraid 50,000 might also not be possible with one PRTG installation. We have one test-setup with 24,000 and that seems to be the maximum for one PRTG installation. The limit here is that PRTG is a 32bit application and with so many sensors it hits the 4GB limit of 32bit applications.

best regards.


Jun, 2011 - Permalink