These are not as much external nodes as they are external services managed by 3rd parties - like Amazon AWS or 3rd party FTP sites for content sharing with partners and clients. Being a sysadmin, I want to know the uptime of these 3rd party services for SLA and preventive trouble-shooting purposes, so I monitor these services via e.g. TCP Port 21 on a DNS name of that FTP site. Ideally whatever port / handshake the service uses. SNMP is out of question of course.
Now, say, it went down, TCP port 21 is no longer open on that FTP site. Did it go down because...
- the host disabled FTP service or changed the port,
- because the host itself went down, or...
- because the site is unreachable for other reasons, e.g. no network route?
This is where ICMP could help (if allowed): if the site still pings then it's (1) more than likely, and this saves a lot of time potentially. If you have 40 such sites (like we do), this could become interesting... Although NPM apparently doesn't support DDNS... so reason (4) could be, "Solarwinds bug".
I might try HolyGuacamole's VB script one of these days.
So this isn't just about ICMP being less traffic, it's more, what options are available to me to check on external service availability, and if the service is down, what options would help me zero down on the root cause?