Hi,
Perhaps this already exists or its not possible for whatever reason(s).
I just had an idea pop into my head for a "reverse" ICMP sensor. Instead of the PRTG server probe sending an ICMP request to a target, the server/probe listens for ICMPs from a host.
One of the biggest problems is monitoring devices that are protected and/or hidden by firewalls, NAT, etc. To overcome this problem I understand Paessler will be beta testing a probe that can live within those protected networks and that don't rely on an expensive Windows Server.
My idea isn't as advanced as the new probe. This assumes the host can ping the PRTG server and that a script or cron job or similar can be run on the host. Essentially the PRTG server would be configured to listen for ICMP on a specific port. A script is installed on a host to send an ICMP to the PRTG server every x seconds. Just like the existing PRTG ICMP sensor, if an ICMP request is not received the host can be presumed to be down.
Multiple hosts could use the same target IP and port and be identified by source IP. If multiple hosts are behind the same public IP, they could be distinguished by sending unique payload sizes or setting various combinations of ICMP flags. There's no need for PRTG to respond to the ICMP, so the host could send a custom UDP datagram instead.
There are obviously limitations to something like this and the remote script can't be managed centrally via the PRTG interface however I thought it might still be worth exploring.
Add comment