Priority of host versus services

[Mike Machado]

On Mon, 2003-02-10 at 13:22, Syed Ali wrote:
> To my knowledge:
>  
> Every defined host in hosts.cfg should have at least one service in
> service.cfg for things to work well.
> The check-host-alive in hosts.cfg does a ping to see if the host is
> alive.
>  
> If the service defined in services.cfg is check_ping, then
> check-host-alive in hosts.cfg and check_ping in services do the same
> thing, right?
>  

If you look at the stock check-host-alive, it only sends one packet and
checks for a specific delay time. What we do is have a check_ping
service called 'PING', where it sends 4 packets that way I can get
warnings based on packet loss and not just an up or a down like the host
version check-host-alive. We also tighten up the delay times so if there
are abnormally long ping times, we get warned as well (possible DoS
attack).

> So if a service check for host A for check_ping is for 24x7, then I will
> get notified 24x7
> However, if the host actually goes down, then I will be notified only
> during work hours?
> That makes no sense...

It does not have to be configured that way. Change the host
notifications to 24x7 as well.

-- 
Mike Machado
mike at innercite.com
InnerCite Inc.
Engineering Director / CTO



-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com