A typical monitoring software normally gives you an option to make a boolean check for some condition, like "is a port available" or "does script X fail".

However, computer networks introduce myriads of conditions which can be complex to overview.

In a graph view, you could see for a deployment model high-level information encoded as colors:

  • endpoints A and B, A and H can "see" each other according to a recent check as configured in the model which can be complex (for example, A can access one set of ports on B, B can ping A, and bidirectional throughput and latency fulfill your SLA for QoS)
  • there are QoS problems in the right part of the graph including nodes B,C,D,E,F
  • Nodes H,D,G represent a problematic area
  • Edges AG and GF have not been checked for more than some time interval (like, you have set up to check every hour which gives green colour, and after 15 minutes it's black).

If you hover/click the graph, you zoom into more details.

The network from the monitoring module even does not have to be a computer network in terms that same software might have further use cases to be reused. For example, you could think of:

  • rendering a service mesh;
  • rendering a micro service architecture;
  • rendering a artificial neural network to display active synapses;
  • rendering public transport network.



You could try ntop and ntopng. I think this tool can discover nodes (given the hosts support LLDP) and learn the structure of the network. (I have tried it many years ago - no idea how good it is now)

Also, some of the visualization-views of the Web_backend of the monitoring tool Nagios Core, or one of its many forks, can give you such network graphs. But there you must enter all rules to check host-availability and service availability yourself.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.