Hi On 2010-02-01 23:32, Cyril Bonté wrote: > This would give something like : > - UP : the server is UP because the tracked server is UP > - DOWN/DOWN : the server is DOWN because the tracked server is DOWN => the > line is in red > - DOWN/MAINT : the server is DOWN because the tracked server is in > maintenance => the line is in brown > - MAINT : the server is in maintenance (no need to know the tracked server > state) => the line is also in brown
In my world, a unconditional down of any server is considered a major problem. Your proposal fits right into that view. So a DOWN is still a failure of some sort while DOWN and DOWN/MAINT show the servers state cleanly. The DOWN/MAINT is actually way better that MAINT alone as it states the cause of the MAINT mode more boldly. > Maybe the DOWN/DOWN can be optional to not break current monitoring scripts > implementations (only DOWN will be displayed). I think, this should be just plain DOWN to keep current behaviour and boldly state the error condition. > What do you think about that ? Me likes :) > If it's ok, I can work on a patch for this tomorrow evening. Great. Thank you for that! --Holger
