I've been reading the blog, and trying to understand the high-level view of how AMP works. Good stuff in the blog by the way!
So, sorry this question is pretty basic. What I gather is that if a service drops out of the netstat listener table, AMP will catch that, or if a listener process dies, AMP will catch that too. But do the Minions make periodic test connections to services to see if those services are still actually available and answering incoming connection requests? What I'm thinking of here is if a service process is still up and running, and it's listener port info is still listed in the netstat table, but due to a bug, or misconfiguration, or whatever reason is not actually answering incoming connection requests anymore, do the ring neighbor minions notice? I don't mean to sound nit-picky at all, I'm just trying to understand what AMP is actually doing. :-) Maybe in practice services don't really go "zombie" like that without also dropping out of the process table and/or netstat available services table, etc.? Sorry if I'm using the wrong terms and names for things here. The story in the blog about reallocating staff to fix bugs in the pbx project before the impending release was cool. Nice solution. :-) --robert _______________________________________________ Assimilation mailing list - Discovery-Driven Monitoring [email protected] http://lists.community.tummy.com/cgi-bin/mailman/listinfo/assimilation http://assimmon.org/
