check_rabbitmq_cluster thresholds should be inverted

[abompard]

Currently the warning and critical threshold trigger when the number of nodes is above the number set in the threshold. They should trigger when the number of nodes goes below the threshold.

# ./check_rabbitmq_cluster -w 2,2 -c 1,1
RABBITMQ_CLUSTER CRITICAL - nb_running_node CRITICAL (3) nb_running_disc_node CRITICAL (3), nb_running_ram_node OK (0)
# ./check_rabbitmq_cluster -w 4,4 -c 5,5
RABBITMQ_CLUSTER OK - nb_running_node OK (3) nb_running_disc_node OK (3) nb_running_ram_node OK (0)

[DVD-CZ]

Hi, after reading table 3 here https://nagios-plugins.org/doc/guidelines.html#THRESHOLDFORMAT it is clear why it behaves the way it does. And it also suggests how do it in a way it could be useful.

I cannot work with perl, and do not know if its possible in some more elegant way, but i assembled it to work:

REM: $critical = $critical{$metric} if (defined $critical{$metric} and $critical{$metric} ne -1);
ADD: $critical = '' . $critical{$metric} . ':' if (defined $critical{$metric} and $critical{$metric} ne -1);

And the same for warning

Or maybe more universal way - pass parameters that already are in such format, like "-c 1:,1:"