Nagios Polling delays during peak hour

[astrosaurabh]

Hi,

We've a Nagios core 3.5 version polling for services every 5 minute interval. Unfortunately, since this morning there's a delay of approx. an hour in polling services.

Could you please assist in troubleshooting this issue..?

Kr,
Saurabh LAHOTI.

[scottwilkerson]

How many hosts and services are on this server being monitored?

One thing I will mention is that Nagios 4 is WAY more efficient that 3.x was.

[astrosaurabh]

Hi there are 457 hosts and 7698 services are getting monitored on this nagios server. I have observed that this delay of check is happening in the peak hours, in non-peak hours check is happening in exact 5 mins which is default one.In logs i can see there are some LDAP contact error. Will LDAP error cause a delay ?

[scottwilkerson]

Will LDAP error cause a delay ?

Generally I would say no, however if you has a lot of plugins that either connected to LDAP, or connected to servers which have to make LDAP connection and it was failing, this could cause the plugins to run all the way until they timed out which in turn would hold a fork open for the whole time, preventing other checks from running.

This again is one of the improvements I mentioned in Core 4. However, if LDAP is in fact erroring, you may want to investigate that as well.

[sandhya12]

we have checked logs found "
[1542888421] Warning: Contact 'linux-dwh' service notification command '/usr/local/nagios/libexec/pnp4n_send_service_mail.pl -p "Sungard: call frontend" -n "$_SERVICECOMMENT$" -a "$_SERVICEAUTHOR$" -c "" -f graph -u > /tmp/pnp4n_send_service_mail.log' timed out after 30 seconds
[1542888421] SERVICE NOTIFICATION: ww_technical_support;lonshbgcoh01a2;COH Cluster Size;UNKNOWN;service-email-graph-int-en;JVM cannot be reached.
[1542888422] SERVICE NOTIFICATION: ito_gdci_kuoni_middleware;lonshbgcoh01a2;COH Cluster Size;UNKNOWN;service-email-graph-int-en;JVM cannot be reached.
[1542888453] Warning: Contact 'ito_gdci_kuoni_middleware' service notification command '/usr/local/nagios/libexec/pnp4n_send_service_mail.pl -p "Sungard: call frontend" -n "$_SERVICECOMMENT$" -a "$_SERVICEAUTHOR$" -c "" -f graph -u > /tmp/pnp4n_send_service_mail.log' timed out after 30 seconds
"
Getting time out error will this cause delay?

[scottwilkerson]

It could if there were many, many of them, but if it is just a few, it seems unlikely to delay normal polling.