Should Active Host checks be disabled?

[Darren]

Hi, quick question around host checks.

We are running nagios core v3.2.3, and on our main nagios server we are actively checking services as follows

Time Frame Services Checked
<= 1 minute: 25 (0.4%)
<= 5 minutes: 5424 (83.8%)
<= 15 minutes: 6463 (99.8%)
<= 1 hour: 6473 (100.0%)
Since program start: 6473 (100.0%)

For hosts we are actively checking as follows
Time Frame Hosts Checked
<= 1 minute: 49 (4.7%)
<= 5 minutes: 1048 (100.0%)
<= 15 minutes: 1048 (100.0%)
<= 1 hour: 1048 (100.0%)
Since program start: 1048 (100.0%):

No passive checks.

I came across notes in another forum about disabling Active Host checks (http://copilotco.com/mail-archives/nagi ... 00517.html) if I have read it correctly that is?

no active host checks

just active service checks (few passive too)

To disable change the host definition/template option active_checks_enabled 0

Host checks are done on demand when any service returns a critical
state (maybe a non-OK state I can't remember). This is default nagios
functionality.

The update of your host status check results will only happen when a
host check is intiated. If a service on that host never goes into a
non-OK state it will never execute the host check. If it does it will
execute a host check, if the host is down it will stop executing
service checks, and use your host check_interval to continue executing
host checks, when the host returns to an OK state service check
processing will begin again.

Does the above behaviour only happen when active host checks are disabled? I am trying to improve the performance of the system as I am aware that we have 6000+ service checks going on and over 1000+ host checks.
We are more interested in the state of the services on the host, but if all of the services go down then it makes sense to mark the host as down and don't bother checking the services again until the host state changes.

Slightly confused how we go about achieving that sort of checking, or if we even need to change anything in our config? Should we disable active host checks? if so what would be the consequence of doing so?
Seems silly to me checking the status of the host, when some services return ok, it only worth checking the host when all services are down per host?

im new to all this nagios monitoring and the guides ive read so far are confusing things more

[sreinhardt]

Honestly, while default host checks such as ping are of limited use, I would suggest against disabling host checks. This can lead to many more notifications if the host is down and services go down, opposed to just getting a host down notification. If I had any one suggestion for you, upgrade to 4.0.3, it has some pretty massive speed upgrades that I'm sure you would benefit from. If I had a second suggestion, look into mod_gearman, it is a distributed way of checking, allowing you to add more workers as your load increases, while decreasing the load the actual nagios process does.