Server reboot notification.

[pacmag]

Hello Nagios Support,

I'm completely new on nagios. We recently purchased the product because it has the capability of monitoring the applications. we had been using an alternative product for long period of time. Our environment is mixed(Windows,linux and routers/switches from different vendors).

In our previous monitoring system, we used to get notification while a server reboots but I couldn't find the way configuring it in our nagios xi server. I have downloaded the check_uptime.pl plugin but not sure how to integrate it with the monitoring wizard? I want to integrate it with Windows SNMP and Linux SNMP wizard so that we can select the option while adding a new host using monitoring wizard.

I've found nagios is not alerting when the server reboots if the server reboot time is less than 5 minutes.

Also, how do I configure the monitoring wizard to check some services in lower time period? I want to chanage the default time for some servers.

Thanks,

[slansing]

This reboot feature you had in your previous solution was presumably just a command that you ran and it emailed the contacts associated with that device with a special message. You have the ability with Nagios XI to do this a few ways, first, you could set downtime for the host and it's services, this would place a comment on the objects, send out a downtime alert, and then cease to alert anyone until the period was over if it was still in a critical state "off the network due to loss of power."

You can also send a custom notification through Home > Hosts/Services > "Name" > Advanced tab > Send Custom Notification. This will allow you to notify anyone who would normally be notified about changes to that object, with a custom message. The reason Nagios XI did not alert you for the host being down is likely 'as you said' it was not checked during that time. You can modify this through a re-running the wizard that was used on that host, depending on the wizard the step where you would change this would likely be step 4 or 5, and you would be changing the check interval. This can also be done through Configure > CCM > Hosts/Services > "name" > check_settings.

[pacmag]

Your first solution works when we have a scheduled maintenance on the server. But in some cases such as our servers reboot automatically after the windows update once we release the updates. In that case, we would only know if the the server is properly updated or not using the notification.

"check_uptime.pl" plugin has the better way of generating alerts if the server reboots/down. I want to integrate this plugin with the existing monitoring wizard(such as Windows SNMP), could you please suggest a solution for this?

You can modify this through a re-running the wizard that was used on that host, depending on the wizard the step where you would change this would likely be step 4 or 5, and you would be changing the check interval. This can also be done through Configure > CCM > Hosts/Services > "name" > check_settings.

Using the above idea, it applies for all the services we choose there. I want different setting for different services such as. I want to check ping in every minute and every 5 minutes for disks. I know we can manually customize the setting for each services after adding the host. But doing this job for 100+ servers is not easier. Is there any way to customize the wizard so that it goes 1 minute for the ping and 5 minutes for the disk using windwos-snmp-monitoring by default?

[slansing]

Well you can use the Windows SNMP Monitoring Wizard to configure SNMP checks on your systems, you would initially need to modify those services by hand, but you can then apply them to other hosts, or even edit the templates the wizard uses, "to check what these are, run the wizard, then enter the CCM and check the service's templates list in it's configuration." That way, you do not need to re-run the wizard, and modify each service that you create, you will have a group already set up.

[vAJ]

Yeah, the reboot alert is based on their db field of "LastBoot" which is a calculated value from the standard SNMP OID "sysUpTime":

Code: Select all

Object	sysUpTime
OID	1.3.6.1.2.1.1.3

I'm not sure how you would get Nagios to perform the same calc and change alert, but it would be useful.

On the other hand, Nagios (with NSClient) gets the Windows uptime WMI value with check_xi_service_nsclient (check_nt) with "UPTIME" as arg2. Problem is, check_nt doesn't allow you to set integer values for warn & crit...

[abrist]

I'm not sure how you would get Nagios to perform the same calc and change alert, but it would be useful.

You would most likely have to write a custom wrapper script to check the uptime oid and alert based on it.

[pacmag]

To be honest, I'm not looking for any special scripting or any advanced configuration all I want is receive notification while the host reboots/goes down. For example.

While the server goes down::

***** Nagios XI Alert *****

Nagios has detected a problem with this host.

Notification Type: PROBLEM
Host: ----------------
State: DOWN
Address: ---------------
Info: CRITICAL - 10.26.240.98: rta nan, lost 100%
Date/Time: 2014-01-23 12:31:53


While the server comes up:


***** Nagios XI Alert *****

Nagios has detected a problem with this host.

Notification Type: RECOVERY
Host: ----------------
State: UP
Address: ---------------
Info: OK - 10.26.240.98: rta 0.320ms, lost 0%
Date/Time: 2014-01-23 12:33:31

We are running most of the servers on VM environment and most of the server come back within a minute of reboot, and mostly we are interested monitoring Disks and hosts(reboot/down). With the default configuration it doesn't notify while the server reboots.

Please help.

[abrist]

If your servers reboot between checks, this will be hard to do with just standard checks. You could decrease the host check interval to like 30 seconds or less, but this may lead to unwanted load/latency if your environment is very large. You could use a passive check that is run on server boot and use that to trigger an alert. Or create a wrapper script to parse uptime and then submit a passive check with a hard state down.

[pacmag]

I'm using the monitoring wizard for adding new hosts for SNMP monitoring. while running the wizard in the step 4 it gives the option for selecting the time period and the minimum time i can select is a minute.

Could you please let me know how do I minimize the time to seconds( as you mentioned like 30 seconds) or so?

You have mentioned the host check, is there particular wizard that does the host monitoring? as I mentioned earlier all we need is monitoring Disks and hosts(reboot/down).

Let me know.

thanks

[lmiltchev]

Could you please let me know how do I minimize the time to seconds( as you mentioned like 30 seconds) or so?

As abrist mentioned, if you have a large environment, this may increase the load on your server, and create unwanted latency issues...

You have mentioned the host check, is there particular wizard that does the host monitoring?

You can probably use our Generic Network Device Monitoring Wizard. All you need to enter is the IP address of the device that you want to monitor. The wizard will create one host and one service check (ping). You can delete the service check later, and leave the host check only. You will be notified on down and recovery states.