Host and service check specific timeouts

[inqrphl]

This PR adds timeout overrides to host/service checks.

Previously, host and service checks each used a global timeout. Users could not change this timeout per host or (more likely) per service. The defaults were 30 seconds for host timeouts and 60 seconds for service timeouts.

Add a field to host/service definitions named 'check_timeout' , which is used as an override.

src/naemon/xodtemplate.h/c

add field 'check_timeout' & 'have_check_timeout' to xodtemplate_host_struct & xodtemplate_service_struct

xodtemplate_begin_object_definition is used to construct the template , set check_timeout to the global host/service timeout values by default.

src/naemon/objects_service.h/c and object_host

add field 'check_timeout' to the service definition, and add the accompanying 'has_check_timeout' that other fields have. Do not exactly understand what it is for, and do not access it in the rest of the PR.

print 'check_timeout' on the fcache_service/host as well

src/naemon/checks.h

add field 'timeout' to the check_result. This is not necessarily needed for logic, but I think that is a good addition for two reasons:

• First, it already has a flag 'early_timeout' that will be set if the check timed out. update_host_state_post_check function checks if it timed out, but not with which value.

• Second, the timeout used might have custom values now. It does not necessarily have to be either of the two global values: service_check_timeout / host_check_timeout. It can be directly saved as a value, taken from a template or a macro etc.

check_result is used in handle_async(host|service)_check_result functions for the broker_(host|service)_check functions, directly getting the saved timeout from check_result might be useful for some use cases. Maybe thruk can use it somehow? @sni

src/naemon/check_host / service

on the run_async_(host|service)_check instead of using the global values, just use the check_timeout from the service/host check now. Does not matter if it was unchanged or not after it was initialized with the global value.

there is also a check_for_orphaned_hosts handler, where the timeout has to be determined again during iteration of all hosts.

src/naemon/commands.c > service_command_handler

no changes here. not registering any commands here to change check_timeout dynamically.

src/naemon/xrddefault.c > xrddefault_read_state_information

no changes here either. this is responsible for printing modified attributes, and we do not modify check_timeout during runtime with a service command

For maintainers:

• Update information in the naemon docs, if required I can also do that. https://github.com/naemon/naemon.github.io
• Overriding the default value is an optional functionality and the default behavior does not change. But there might be users who put check_timeout in host/service definitions, and they will be effected accidentally. Have to warn them in the changelog.

I can sign the comments if necessary. Please quash before merging.

[sni]

Looks good. Could you please have a look at the failed test case. Probably just missing the new values. Also it would be nice to add a Changelog entry in the NEWS file and also adding the new attribute to the docs on the website.