In huge systems, there are many items to monitor. It is hard to find monitored parameter to add to graph and it is sometimes slow. Next problem is that sometimes wee need to compute graph items by formula. Maybe it would be better if there is no selectbox but inputbox. I can write any expression to graph.
eg.
{server:netloadin[eth0].avg(120)/server:netloadin[eth1].avg(120)}
would monitor firewall with twoo interfaces eth0 and eth1 . If number is close to 1, everything is ok, another value could be some attack and would be nice to see it in graph.
or
{server:diskfree[/].change(0)} can be used to see changes in disk space.
eg.
{server:netloadin[eth0].avg(120)/server:netloadin[eth1].avg(120)}
would monitor firewall with twoo interfaces eth0 and eth1 . If number is close to 1, everything is ok, another value could be some attack and would be nice to see it in graph.
or
{server:diskfree[/].change(0)} can be used to see changes in disk space.
Comment