[Observium] Verifying timeouts

We've been using observium for well over a year now and extrememly happy with it. We have devices all over the world and recently started to see some strange reboot, high cpu loads and device up/down messages when the devices appear to be ok I recently noticed it seems to be the ones further away.

I noticed the log really doesn't specify anything specific and my php skills are not up to snuff so I was wondering how would one go about verifying if we are seeing timeouts on the polling to know the best way to rectify the situation we are experiencing. Can someone verify is say the timeout of a cpu poll would actually how max cpu load?

Thanks, Bryan