<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000099"> Marc Powell wrote: <blockquote cite="mid:09B6D808-6A50-4F19-BEF0-300AE3F9B8B6@ena.com" type="cite"> <pre wrap="">On Jun 1, 2009, at 3:09 AM, Arnar Þórarinsson wrote: </pre> <blockquote type="cite"> <pre wrap="">So there is no way of getting just the host down alert when a host goes down ? To explain a little, lets say I'm monitoring CPU, memory and disk space on a host. The host goes down and Nagios sends an alert by email for the host down event and also for the CPU, memory and disk space events. All I need to know about this event is that the host is down. Just think that it's not neccessary to send an alert email about services on a host that is down. </pre> </blockquote> <pre wrap=""> And so does nagios. As I said earlier, nagios does this automatically. To restate - when a host is down, nagios suppresses all e-mail notifications about that hosts services, but will still display them as down in the GUI. It will only send the host down notification. The first section of <a class="moz-txt-link-freetext" href="http://nagios.sourceforge.net/docs/2_0/networkreachability.html">http://nagios.sourceforge.net/docs/2_0/networkreachability.html</a> states it best. It still applies to 3.x but I haven't found the section that states it as clearly. -- Marc </pre> </blockquote> If I'm interpreting your question correctly, you're saying that when one of your servers actually goes down, you ARE getting alerts (email/SMS/whatever) for more than just the host being down??? I see what Marc's saying... he's telling you this shouldn't be. Nagios was built to check first that the host is up and reachable, and if its not to notify you the host is down, but to not ALERT you about all host-dependent tests that are now failing. Nagios will still try all tests and fail on them and the web interface will reflect more than just the HOST DOWN, but the only email/SMS you get should be for the HOST DOWN. However, you may need to clarify what you mean by *down*. *Down* does not always mean off or 100% non-responsive. In the case of *nix systems I've seen quite a few times where a server will hang, fail, or segfault but still be reachable over the network. The reason is that parts of the OS are in memory and things like pings from remote hosts still respond, even though the overall functionality of the host itself is down (ISP's get this a lot: host pings, but you can't ssh in, for example). If Nagios can ping the host, it will then try the other tests and alert on them. Here's a quick way to narrow this down: turn off the server (shut down and pull power). The Nagios web interface should show the host down and all tests as failing, but the only email/SMS you should get is the host down. If you still get emailed/alerted then you might have a configuration error. Perhaps you didn't properly define your host checks as opposed to service checks? Do you have a check_ping or check_icmp host check for each host? AD </body> </html>