A further note: Also if you use an external monitoring service(e.g. Pingdom, Wormly, Exactstate) that performs DNS traversals for each check instead of relying on a recursive server, this likely explains any increase in false alarms.
Yeah, after I read your comment I recheck my Pingdom's logs, those weird timeout errors since one year ago are now gone (I've switched to CloudFlare). Back then I was scratching my head wondering why Pingdom generated so much errors from random different locations while people reported that they were able to access my sites jut fine. Thanks!
