Type: New Feature Request
Affects Version/s: None
Fix Version/s: None
Component/s: Server (S)
this feature request is meant to be a continuation of
ZBXNEXT-341 which seems to be ended up with implementation by internal events.
Internal item zabbix[host,,items_unsupported] is IMHO golden mean between overall zabbix[items_unsupported] (which is not specific enough) and internal events for unsupported items as implemented in
ZBXNEXT-1575 (which are too specific).
What is IMHO missing, is analogous internal item for unknown triggers, e.g. zabbix[host,,triggers_unknown], because there may be cases, when all items in trigger expression are supported, but trigger is still in unknown state (e.g. disvision by zero).
Internal events may be useful for e.g. some debugging, but in production use, there are issues:
- internal events do not have event-representation in Monitoring > Problems, therefore frontend user is unable to display or filter them
- when notified, notifications about individual internal events may cause in some cases serious notifications flood (I personally experienced the case, where it knocked to knees e-mail infrastructure of the organization for several days)
- this detail - individual event for each trigger - is IMHO not needed much, because when detected, it is easy to filter out unknown triggers via frontend, so it is not needed to be overloaded by many individual events, when the main problem is: "monitoring of host XYZ is not 100% functional"
I have a golden self-monitoring rule, having trigger for non-zero zabbix[host,,items_unsupported] on every host in every ZBX instance. It would be fantastic, to be able do the same also with zabbix[host,,triggers_unknown].
In case of a positive decision on this humble plea, thank you in advance.