[Nagiosplug-devel] check_http plugin suggestion

J.T. Moore jtmoore at autosportcatalog.com
Thu Oct 27 15:08:12 UTC 2005

    It appears that the check_http plugin ignores the  --string option 
    when the http status code is 4xx
    Normally if the string is not found, the plugin sets the service state
    to critical, but if a 4xx status code is returned, the service state is
    set to warning even if the string is not present.
    I suppose there may be times where this logic may be preferred, for
    example a http 200 status code with different content would indicate 
    a likely defacement and is a critical problem where as a http 400 
    status code would indicate a less serious service disruption or 
    configuration error
    If people are actually using the above logic, it shouldn't be changed,
    in my case it would be nice to add a option to always compare the string
    value and always treat the string not being found as a critical error.
    What I am doing is using Nagios and a few other tools to set up a
    for a fault tolerant web farm. To me a warning is, the page look longer
    load than is desirable (as specified with the -w argument). A critical
error is not
    returning a page with the specified string (specified with the -s
    before the critical time (specified with the -c argument). Ideally, I
    like to write an event handler that wouldn't do anything about warnings
    in a soft state, but that would stop sending traffic to a server if it
was in
    a hard warning state or any type of critical state.
Thanks for the cool tools,
J.T. Moore
International Auto Parts
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nagios.com/pipermail/nagiosplugin-devel/attachments/20051027/04319b5b/attachment.html>

More information about the Nagiosplugin-devel mailing list