Interesting approach.

What we’ve done is to write an app that runs on a separate machine that simply does some asterisk -rx calls to the running Asterisk instance via an SSH library and then evaluate the string returned.

For example, to monitor our registered SIP service providers, we compare the output for the sip show registry CLI command when the system is in the correct state, to the output we get on timed checks every five minutes.

If a SIP trunk goes down, the just-obtrained string does not match the stored, correct string anymore, and our app starts sending SMSes / raising alarms.

Been running for about three years now and works suffieciently for our widely separated branches all over South Africa.