I really do understand this, on a deeper level than might be imagined ;). I have no vested interest either way personally, I’m just offering some unsolicited thoughts...
I always looked at the “ping only” topic like asking a 10lbs. (4.5kg or 0.7149 stones) sledge hammer to drive a nail. Sure, it can do it but it’ll probably be messy and ugly because a sledge hammer was never meant to drive nails, and when you use tools in ways they weren’t really designed for, things can get messy. Just use a hammer instead.
I have done three things to achieve icmp only monitoring and also stick within the base code (as a developer, I hate mucking with base code or re-inventing wheels when I want to do different things). Icmp-only monitoring to me, always just seemed silly (not in purpose) to try and work into an existing and established application that didn’t already do it.
It’s just too easy of a problem to solve otherwise, than to go to the time/expense of refactoring an application IMHO.
- I use RTT SLAs to monitor next hop devices (which OBS is fabulous at picking up), assuming the device hopped from is SLA capable and OBS monitored.
- I use Fping (Smokeping) integrated with OBS. Using the Smokeping web wrapper makes a great RTT only tool, and you also have the OBS integration for when it makes sense to use on OBS devices.
- I have a small (very small) bash script that reads a text file of IP address and launches one thread per device to run continuous pings. Not a very scalable script, but like you, I only need it for a very few limited cases. The script uses sendmail to email me if it can’t reach something.
Thanks,
Ryan