I'm looking for a (or multiple if needed) system with web interface to monitor a number of remote networks. All of these networks contain multiple clients, servers, routers, switches, printers, ...
Some info we would like to gather is the remaining ink supply in printers (via SMNP), checking the status of clients (remote assistance via VNC, RDP,... is a plus), monitor network load, check if services are still up, be notified if anything goes down... You get the picture.
The servers are mixed Windows and Linux (CentOS and Ubuntu), clients are mostly Windows.
So far, I have already investigated some possible solutions:
Can anyone recommend one of these or suggest something else? Does anyone know of a good implementation for Nagios (like Centreon)?
One solution for all of these problems would be great.
Nagios is all about alerting you when actions need to be taken, like printers need to be filled or hosts are down or their loads are too high, or drives have fallen out of their RAID array... It's not just an "install and turn on" sort of solution though, you will need to spend time creating the monitoring configuration, but if you have a lot of identical hosts it may not be as big a job as if they are all different.
Cacti is more about making graphs, like amount of traffic on a network link, or CPU utilization on a router or host, to get long-term graphs of utilization trending and capacity analysis. Though for hosts we prefer to use munin, because it's pretty much self-configuring. Cacti, like nagios, tends to use need a bit of setup. Again, unless many of the devices are identical.
Both of these, of course, are a process and not just a fire and forget solution. Expect to regularly do updates and tweaks.
But, we use nagios for monitoring of things that need immediate action, and cacti and munin for capacity analysis.
Solarwinds NPM is a Network Monitoring System that can monitor switches, routers, printers, and workstations with SNMP in one dandy web interface.
With their add-on product Orion Additional Polling Engine you would be able to monitor several remote networks.
There is a handy toolbar on all of the "nodes" that allows you to quickly open up a telnet, ssh, or remote desktop session but does not tunnel traffic, which may pose a problem on remote sites.
I find their built in alerting to be very helpful. I have setup alerts like you have described for situations like our bonded T1 line's utilization reaching a certain percent or a server going offline.