On Wed, Aug 28, 2002 at 09:42:26AM -0500, Allyn Baskerville wrote: > I use rrdtool with a custom front end to keep historical data on key > parameters for servers, routers, and switches, and then mon to alert me of > all outages. This seems to be a good combination, as the first provides me > with device health and troubleshooting data (and often times I can see > impending network outages), while the later notifies me of the actual > outages.
This is what we do too. We've built host monitoring and DNS monitoring (http://www.campin.net/DNS/graph.html) from scratch using RRDtool, and use scattered Mon installs to monitor different datacenters. We use Mon to fix downed IIS processes and also some commercial software that freezes. Mon's flexibility is key for enterprise monitoring. We have a NOC that watches openview and receives snmp traps from various bigbro installs, but our Mon installs do the job of a room full of people. We our problems fix themselves, no one has to stare at a screen 24x7 and call people. We don't even use the NOC, we don't need them. I'm working on a sysadmin book that has some promise of being published, and Mon is all over it. -- "Truth is the most valuable thing we have. Let us economise it." - Samuel Clemens
