Latencies increase drastically after 33 hourse of uptime

[dnelson]

Hi abrist,

Thanks for the link. The link you provided appears to be for Nagios 4.x and differs slightly from Nagios 3.x. I'm including both here to avoid any confusion.

Nagios 3.x large installation tweaks: http://nagios.sourceforge.net/docs/3_0/ ... weaks.html

Nagios 4.x large installation tweaks: http://nagios.sourceforge.net/docs/nagi ... weaks.html

Regards,
David Nelson

[abrist]

The link you provided appears to be for Nagios 4.x and differs slightly from Nagios 3.x.

True enough. Both versions do limit the environment variables, so make sure that your checks do not use them before you deploy the changes to production. (most plugins do not use them anymore, but there are some older, heavily customizable notification addods that do . . .)

[dnelson]

On the advice of an individual that was experiencing the same issues as I was, I disabled the embedded Perl interpreter (set 'enable_embedded_perl=0' in nagios.cfg). Restarted Nagios and waited. It's now been 5+ days since doing this and service and host latencies are stable in the 200 and 400 mS ranges, respectively; the 1 minute load average mean is 1.78 with the max being 2.48; kernel/system time is stable at around 12%.

In summary:
- Large installation tweaks enabled (use_large_installation_tweaks=1)
- Free child process memory commented (#free_child_process_memory=0)
- Child processes fork twice commented (#child_processes_fork_twice=1)
- Disabled the embedded perl interpreter (enable_embedded_perl=0)

Thank you for all your suggestions and ideas.

David Nelson