Check_MK and distributed WATO

This is not yet an article yet (wink) 

It will become one as more questions come up.

17:22 < midoski> darkfader: have you ever done the WATO multisite stuff?
21:21 < darkfader> midoski: yes
21:22 < darkfader> i had one server per "colo" myself, and bigger setups when i
still did that as a job
21:22 < darkfader> now i went back to just one nagios box because it can handle
it easily
21:23 < darkfader> the downside is i can't see "what happened" after the fact if a line goes down or similar
21:29 < darkfader> what's important... hmm. you normally want to have a
"master" site that doesn't do much monitoring and is used as
the point of configuration for all slaves
21:30 < darkfader> basically it should only monitor those other sites, possibly
also do BI aggregations since those usually will span over
multiple sites
21:30 < darkfader> and you need to invest a few hours to have a really nice
wato folder structure
21:31 < darkfader> i often tried to convert something existing (or was asked
to) and those setups never came out to a standard that I was
satisfied with
21:31 < darkfader> if you design it from scratch it will be much better to
maintain
21:32 < darkfader> that has a risk of "losing" a few custom checks as you move
them over. i made a livestatus query in each site that
listed all services on all hosts and then I was able to
track down / verify all services

  • No labels