If you’re a hardcore Gmail user, I don’t need to tell you that Gmail was down yesterday. The Gmail Blog has an explanation of the outage, which officially lasted for 100 minutes. So, here’s the reason you couldn’t get to your Gmail yesterday, straight from the horse’s mouth: apparently Google took down some servers for routine maintenance. This would have been fine, but because of some recent improvements to the way Gmail handles requests, the request routers became overloaded. This caused them to shunt the workload over to other request routers, which quickly became overloaded, too. Oops.
Fortunately, Google has tons of extra capacity just waiting to be switched on in this kind of emergency, so things were back up and running relatively quickly. Google has quite rightly determined that their request routers need some failure isolation improvements to prevent a similar chain reaction in the future, and they promise they’re working on it. Maybe this will quiet down all of the “#gfail” naysayers who freaked out via Twitter the minute the service went down yesterday.
