Page 1 of 1

Cadence necessary for Service check graphing?

Posted: Mon Mar 18, 2019 1:58 pm
by yo_marc
Hello Nagios support,

I have a user of our XI system who has configured 3 passive checks to receive data once every 24-hours. We notice that the built-in graphing tool was having a hard time plotting the data for these 3 checks. Only one out of 3 service checks are working, the others two were still reporting "0" for the perf-data in the graph. (I confirmed they are getting the data in the check results).

I forced some manual passive-check-results for the two which were missing graphing data, and the graphing kicked in for them. But then we noticed the 3rd working one started to show a gap in the plot-line. I have seen this in Active checks, typically when they are not receiving data for x-amount of time.

In summary, I have the hunch that the 1x/day cadence might be too lengthy for the graphing tool to understand what to do. Would that be correct? If so, is there a minimum cadence that is necessary in order to have graphing recognize the data correctly?

Thanks,
-marc

(XI 5.5.8, Centos 7, all Components at latest rev)

Re: Cadence necessary for Service check graphing?

Posted: Mon Mar 18, 2019 3:16 pm
by cdienger
The default heartbeat is 8640 seconds, meaning the data isn't inserted if there is an interval greater than 2.4 hours between points. You can verify the heartbeat settings on the databases by running:

Code: Select all

rdtool dump /usr/local/nagios/share/perfdata/<hostname>/<servicedesc>.rrd | grep -i heart

Re: Cadence necessary for Service check graphing?

Posted: Tue Mar 19, 2019 9:20 am
by yo_marc
Thank you!

Re: Cadence necessary for Service check graphing?

Posted: Tue Mar 19, 2019 11:11 am
by cdienger
Glad to help! Anything further or can we lock this one up?