If host reboots or goes down suppress service Notifications

[nagiosEngie]

Hello Nagios Crew,
I am unabl to find this info:
is it possible to tell Nagios not to send service alerts when host goes down or reboots.

Can you help?

Thanks
sandro

[scottwilkerson]

This is the default behavior

[nagiosEngie]

Hello,
well then I think I have a problem because when I reboot my hosts all checks on services will give me an CRITICAL event ora an UNKNOWN event instead of jus giving me the host DOWN.

Where is this set?
Do I have to define the hosts as parent of the services associated to it?
if true, How is this done?
Thanks
Sandro

[scottwilkerson]

Can you send you profile.zip form Admin -> System Profile along with the name of the host that rebooted and the service that alerted?

[nagiosEngie]

HEllo Attached the profile.zip
I have also noticed that I never get the hostdown hostup notification.

SAndro

[scottwilkerson]

along with the name of the host that rebooted and the service that alerted?

I need the name of the host that went down

[nagiosEngie]

Hello,
I found out the issue is connected to the check interval we are using which is 5min.
Nearly all servers we have reboot nearly all under 3 min.
I poll (ping for up/down) the server at time T0,
while waiting for the next check at T5, the server reboots at T2.
The nagios server dose a service check (for example) at T3.
The server is still rebooting and I will receive a UNKNOWN notification for the service (agent and Nagios server are not communicating, agent is still not up)
Server starts at T4 (about 4 min to reboot).
I get the OK for all the UNKNOWN notifications.

You can see the sequence in the attached image.
Attached also an updated system profile file.

Thanks
SAndro

[scottwilkerson]

So for many of these services you have set the amount of retries before alerting to 2

this means it checks every 5 minutes + 1 minute for the retries, potentially the service could be sending notifications at 3 minutes in.

I would strongly suggest polling the hosts more frequently (every 5 minutes) and setting max_check_attempts to 5 on the services.

This would force the hosts to have checked at least 1 time before a service notification can go out.

These settings would be the default in the Wizards.

[nagiosEngie]

Hi Scott,
I have reduced the polling time on hosts (ie ping up/down) to 2 minutes and services are checked every 5min. retry is set to 2 and retry time to 1 min.
I am just a bit worried about the 2 min polling on host (ie ping up/down). Can this be too stressful for the server?
We have around 200 hosts with a predicted size of 500. There is arount 9 services per Host ( this is an avarage value)
We have at the moment a VM with 8 cvcpu and 8 gb of ram.
Thanks
SAndro

[scottwilkerson]

Hi Scott,
I have reduced the polling time on hosts (ie ping up/down) to 2 minutes and services are checked every 5min. retry is set to 2 and retry time to 1 min.
I am just a bit worried about the 2 min polling on host (ie ping up/down). Can this be too stressful for the server?
We have around 200 hosts with a predicted size of 500. There is arount 9 services per Host ( this is an avarage value)
We have at the moment a VM with 8 cvcpu and 8 gb of ram.
Thanks
SAndro

These ping checks should not be too stressful at all, they are very light weight. However, even if the hosts were polling hosts at 5 minutes you wouldn't run into the issue.

You should be in good shape, and I doubt you will see any difference in system load