TIME parameter in analysis.cfg
Hi all, Just wanted to check I have my understanding of the TIME parameter on rules in analysis.cfg correct.
If I have the following lines in analysis.cfg
DISK /nmon 95 98 TIME=W:0800:1800 DISK /nmon 95 101
then if I'm correct this will give the 95/98 thresholds between 08:00 and 18:00 on weekdays and the 95/101 thresholds at all other times?
Clarification of this would really help me with some exclusions I need to set up.
Regards, Neil.
On Thu, October 8, 2015 1:37 am, Neil Simmonds wrote:
Hi all, Just wanted to check I have my understanding of the TIME parameter on rules in analysis.cfg correct.
If I have the following lines in analysis.cfg
DISK /nmon 95 98 TIME=W:0800:1800 DISK /nmon 95 101
then if I'm correct this will give the 95/98 thresholds between 08:00 and 18:00 on weekdays and the 95/101 thresholds at all other times?
Clarification of this would really help me with some exclusions I need to set up.
Regards, Neil.
Hi Neil,
This is correct. Rules take into account HOLIDAYS, TIME, and EXTIME values, and one that doesn't match the current datetime will be skipped over when it's time to evaluate colors.
For 'DISK' checks, the first matching rule wins (which is why the 'DEFAULT' section should be at the bottom of analysis.cfg and any analysis.d/* sort of config file loading should be just prior to it) and is considered the "canonical threshold" for that partition.
You can run 'xymoncmd xymond_client --dump-config' and should be able to follow the exclusion logic linearly.
For things that aren't just threshold evaluations (which includes PROC checks), multiple rules can match and all of them are generally applied (with separate sub-status lines). As usual for all xymond_client tests, the worst color within that status is the overall color of the status.
A suggestion was made a while back to allow a 'DEFAULT'-like tag onto individual (non-threshold) rules, which would have them apply only if no other rule had applied yet. Unfortunately, the logic for that would be rather spread out and "intentional exclusions" are tricky to track, so it's a non-trivial logic patch.
HTH, -jc
participants (2)
-
cleaver@terabithia.org
-
neilsimmonds1808@gmail.com