Bandwidth consumption estimate

As the agent performs checks on servers and transmits the results back to the central monitoring server, it consumes only very little bandwidth. The amount of bandwidth required is determined by two factors

• The check interval - the more frequent checks, the more results and therefore more data to send
• The number of active checks - the more checks to send results for, the more data we send

There is a little bit of overhead to establish the agent check-in at the configured check-in interval, but this is insignificant compared to the transmission of the check data in all normal configurations.

Furthermore, if the agents check in to a satellite and the satellite forwards the results to a master, the exact amount of data will also be somewhat affected - but all in all the amount of data in need of transmission still follows the amount of check data gathered.

As a rule of thumb, for checks configured to run on 30 second intervals, you should expect to see
one byte per second per check
being used for agent check data transmissions.

So, if you have 100 agents, each with 100 checks and each check is configured to be run every 30 seconds, the total averaged network data utilization will be roughly equivalent to:
100 [agents] * 100 [checks] * 1 [byte per second per check] = 10 [kB per second]

So a fairly sizable network of servers can be monitored with 100 checks per server using a total of around 100kbps or 0.1% of an old 100Mbit ethernet.