Configuring a red alert for: Machine recently rebooted 🔗 link

I tried to configure an alert that should stay red for one hour after a reboot.
The toolonglimit should be infinite. My analysis.conf contains:

HOST=hal
 ??????? UP 1h -1? red

This leads to a red cpu status:

"Machine has been up more than 0 days"

Obviously the toolonglimit of -1 isn't working as I expected it.

dump-config shows 0 in thetoolong-column:

bbrother at xymon:~/server $ bin/xymond_client --dump-config | grep UP
UP 3600 0 HOST=hal (line: 358)

 ?tested under 4.3.28 (the dump-config test showed the same result under 4.3.30)

Is this the expected behaviour? I worked around it by specifying a lot of weeks for toolonglimit.

Hello,


Did you try this ?

HOST=hal

 ??????? UP????? 1h red


Cordialement,
Damien Martins

Le 25/05/2020 ? 15:44, M. Mueller a ?crit?:

Hello,

It seems that the function durationvalue in timefunc.c does not handle 
negative values. (Which seems pretty reasonable). I'm not sure if we 
ever need negative durations other than -1. Anyway I worked around this 
by adding:

         int sign = 1;

         if( *dur ='-' ) {
            startofval = dur+1;
            sign=-1;
         }else{
            startofval = dur;
         }
to the start of the function and returning:

   result*sign;

Cheers

Martin

Hello,

It seems that the function durationvalue in timefunc.c does not handle 
negative values. (Which seems pretty reasonable). I'm not sure if we 
ever need negative durations other than -1. Anyway I worked around this 
by adding:

         int sign = 1;

         if( *dur ='-' ) {
            startofval = dur+1;
            sign=-1;
         }else{
            startofval = dur;
         }
to the start of the function and returning:

   result*sign;

Cheers

Martin