i'm curious how people combine active and passive checks
together.
typical scenario not covered by the doc:
* SNMP polling of some OIDs
* SNMP traps
As for most active checks, max_check_attempts is set to some
number > 1 to avoid false negatives.
The problem is that when a trap arrives, it generally means
that the service has moved to a new different *HARD* state.
But if you handle SNMP traps by sending them as external
command, then it'll just be the first attempt of many, and
you're in a soft state until the active check kicks in which
really defeats the purpose of setting up traps.
i can think of couple ways to fudge with this:
(1) send a SCHEDULE_SVC_CHECK command, but there's no guaratee
as to when that will happen, so there is still an unpleasant
delay
(2) send the passive check max_check_attempts times
this is rather kludgy, and you need to guess that value.
(also misleads about how many times it really happened:/)
but really and unless i am missing something, i think nagios
needs to be extended to support one or more of the following:
(1) PROCESS_SERVICE_CHECK_RESULT needs a new field to allow
forcing a new HARD status
(2) max_check_attempts needs to be split in 2 for active and
passive
Thoughts? Comments?
Christophe
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: kalt@taranis.org