I have some servers that I need to monitor traffic levels for individual tcp and udp ports, and present as a graph for the admins to visually check. I have tried the regular Quality of Service right menu with tc/fireqos, and while that gets the information out, the graphing is not really what is needed due to the differing levels of traffic (1kbps for most traffic, but a few with 1Mbit or more), and the sheer number of ports.
It would be nice to have a graph per port, with a 24 hour average line plotted. Some kind of alerting for those whose traffic drops to zero over a configurable timeout (I’m thinking 10 to 30 seconds, but …) would be of great help as well.
Please note that this is not a physical port I’m mentioning above, but tcp or udp ports. I expect to define between 20 to 50 ports per server.