This article applies to PRTG Network Monitor 7 or later

The bandwidth used for the core and probe connections depends on the number of sensors being monitored, the number of channels defined for the same, as well as the scanning interval.

Quick answer:

100 SNMP sensors, each with 2 channels (in & out) generate roughly 1 kbit/s plus 20 bytes/minute (due to the core "keep alive" message). This reflects a daily load of about 11 Mbyte/day.

Detailed answer:

A breakdown of the bandwidth required for remote probes and the core server is provided below:

Premise: sensors monitored

• 24 ports * 2 channels (in & out) = 48 values
• 15 server sensors * about 4 channels (depending on kind) = 60 values
• 2 firewall counter * 2 channels = 4 values

Calculation

Let us assume an interval of 1 minute for all sensors and generously round up.

We have to send about 50 messages with 150 values from the probe to the core.

A value uses 28 bytes. A message uses 22 bytes plus the values (if there is no message text).

50 messages * 22 bytes + 150 values * 28 bytes = 5300 bytes / minute = 706 bits/s = below 1kbit/s. Or about 7 Mbyte/day for the probe -> core messages.

The core sends a "keep alive" message to the probe every 60s. That is 20 bytes / Minute.

When the probe connects or anything is changed in the configuration the required information is sent to the probe once. This does not create a constant load.

Increasing the interval from 1 minute to 5 minutes creates 1/5 of the load.

Note: These numbers apply to common monitoring load. For specific monitoring tasks they might differ. Syslog Receiver and SNMP Trap Receiver sensors, for example, generate extra load on the connection when data is requested from the core (like message lists or reports).