Here’s Why Gmail and Google Chrome Had Issues Yesterday
We panicked along with everyone else yesterday when Gmail went down. Without email, how would we find out about all the new T-shirt designs and Kickstarter campaigns that are happening? It was maddening. Sure, the outage only lasted about 20 minutes, but it really messed up a very brief part of our day, and even included issues with Google Chrome, and we want answers. Thankfully, Google engineer Tim Steele has them.
Basically, the outage to Gmail was caused by a misconfiguration of some load-balancing servers. Things like that happen. Who hasn’t misconfigured a few load-balancing servers in their day? What was surprising, though, is that the outage affected more than just Gmail. Many users reported that their Chrome browsers were crashing as well. Chrome browsers aren’t supposed to do that. Chrome kind of prides itself on not shutting down the whole browser when something goes wrong, just the failing tab. So what happened there?
It all has to do with Google Sync, which takes the heavy lifting of computing off the traditional desktop and onto Google’s own servers. The idea being that Google servers are more reliable than the average desktop. The problem is that if my desktop fails, I have a problem. If a Google server fails like it did yesterday, everyone has a problem.
Tim Steele said the error was, “due to a backend service that sync servers depend on becoming overwhelmed, and sync servers responding to that by telling all clients to throttle all data types.” By telling clients to throttle all data types, the problem became more widespread than just one tab or Google service, and the whole browser shut down.
With services like Google Sync, Apple’s iCloud, and Windows Live becoming increasingly popular and involved in our lives, this sort of widespread outage could be more common in the future. The integration of data and services in the cloud being available across our computers, phones, televisions, and more is great when it’s working, but small problems become much bigger issues when they’re shared by millions of people instead of just one user.