We should add some basic monitoring of fedoramagazine and perhaps some http check that makes sure it's serving it's front page at least.
Hi i'm working on this ticket. Could you please assign me?
The nagios host definition under '''ansible/roles/nagios/server/files/nagios/hosts''' could be:
{{{
define host { host_name magazine.fedorainfracloud.org alias magazine use defaulttemplate address magazine.fedorainfracloud.org }
}}}
nagios magazine.fedorainfracloud.org magazine.cfg
The nrpe definition could be (adding magazine) under '''ansible/roles/nagios/server/files/nagios/services''' editing nrpe.cfg file:
{{{ define service { host_name hosted03, bastion02, bastion01, bastion-vpn, sundries01, sundries01.stg, sundries02, wiki01, wiki01.stg, wiki02, pkgdb01, pkgdb02, pkgdb01.stg, magazine service_description nrpe check_command test_nrpe max_check_attempts 2 normal_check_interval 2 use defaulttemplate }
attachment nrpe.cfg
The Http check could be added in website.cfg under '''ansible/roles/nagios/server/files/nagios/services
{{{ define service { host_name magazine service_description HTTP check use websitetemplate check_command check_website!magazine.fedorainfracloud.org }
attachment websites.cfg
Hi what do you think about it? It's ok or needs some modification?
That looks good, but can you make it in the form of a git patch? ;)
Basically you clone the ansible repo on batcave01, then make your changes to your copy. Then do a 'git add' for new files and 'git commit -a' and enter a commit message, then use 'git format-patch' to create a patch to attach here. ;)
This way we can see exactly whats changed and also when we apply it you will get credit, etc. :)
Please ask on irc if that doesn't make sense or you get stuck anywhere.
Add monitoring for fedoramagazine Add-monitoring-for-fedoramagazine.patch
You're right! :) Everything it's clear and i added the patch.
Thanks nirik! ;)
ok, applied! :)
I did have to do a bit of tweaking (which is normal with nagios as it's config is such a pain):
I swapped host and alias, we for some reason use the short name for host and the fqdn for alias.
I added the nagios/client role to the magazine playbook so it could be answering nrpe on its end.
I added the nagios security group to the magazine cloud instance so nrpe would be allowed to it.
I think all is working now. Thank you very much for working on this!
Login to comment on this ticket.