Running commercial version of CheckMK and wanting to fine tune default availability report for a subset of hosts (WAN links monitored with ICMP) with labels (all that already set up and working).
The uptime vs warning vs critical etc is slightly skewed due to reporting indicating warning/critical for losses (this is correct) but losses also occur when either WAN links are saturated and there is no bandwidth left.
So 4 services are being monitored:
A. [red] Host ISP WAN router 1 - ICMP service only (for uptime, latency and losses/RTT)
B. [blue] Host ISP WAN router 2 - ICMP service only (for uptime, latency and losses/RTT)
C. [orange] Client Router Host with a SNMP service that monitors interface bandwidth to WAN 1
D. [purple] Client Router Host with a SNMP service that monitors interface bandwidth to WAN 2
Availibility report looks at services A and B above and reports availaibility on those those services.
I would like to exclude events/avialiblity calculation during periods where service C (linked to A) and D (linked to B) is in warning or critical state.
This would mean when pulling a availability report for A and B, it neglects any input/affect caused by bandwidth constraints from C or D.
Hope this makes sense and will post a primitive diagram to aid in understanding.
I don’t think the availability can give you this sort of dependency logic.
Maybe a BI construct is possible, but I can’t think of one right now.
What might work is setting the service of A/red or B/blue into downtime depending on the state of C or D and then decide that you exclude the scheduled downtimes from the availability’s warn/crit calculation.
So after review of this article/video, it would seem that it will solve the requirement
I also noted that there are several requests to make service dependency an official feature, but no more on that now.
Due to complexities, the regex and fully automated options wont be an option, and I will have to create such a rule for each and every “pair” on each and every device/host/service this is required - or am i missing something?
Herewith my manual settings, but this gives me an error " ! Exception during execution. See details" (Screenshots below)
[under conditions i selected the explicit host applicable (ISP 1 Router)]
the plugin assumes a checkmk instance running behind an apache that serves https.
Do you use your checkmk on docker or only with http?
To verify if that is it, you could use the option “connect to this central site” and insert localhost:5000 ? does that work?
Now… the further work to see if and how this can be wangled to accomodate our requirement to trigger the A/RED host (service) to downtime if the service C/Orange is warning or critical.
Wish me luck an will post progress here in the coming days
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.
" ! Exception during execution. See details" (Screenshots below)
