A major outage of Google’s DoubleClick advertising delivery service hit Websites worldwide, slowing or blocking access to thousands of Websites and the ads displayed on those sites.
Dynatrace, an Application Performance Management (APM) service company, tracked the DoubleClick outages and reported it affected the performance of 55,185 Websites worldwide.
The problems began at about 9:40 a.m. EST, per Dynatrace, and lasted about two hours. Sites impacted include USA Today, the Wall Street Journal, Forbes, BBC.com, YouTube as well as the Guardian and the Enquirer news Websites in the UK.
At 10:28 a.m. EST, Reuters reported that Google disclosed that it was “experiencing issues which are preventing ads from serving. Our Technology teams are currently working to restore service as quickly as possible.”
At 12:27 EST, Reuters reported that Google had released a message that said, “Our team has worked quickly to fix the software bug and DFP (DoubleClick for Publishers) is now back up and running, so our publisher partners can return to funding their content.”
However, not all of the effects of the outage were negative. Some sites, unexpectedly relieved of their usual ad lineup, loaded faster. And some humorous comments circulated on social media. Gannett Co, the parent of USA Today, tweeted: “Enjoy your ad-free experience on http://usatoday.com right now, brought to you by Google!”
Data from Dynatrace shows that almost every industry was impacted by the outage, including travel, news, entertainment and retail sites.
“It’s rare for an organization of this maturity to suffer a major outage. These people are really good,” David Jones, director of Sales and APM evangelist at Dynatrace, told eWEEK. “We’ve seen nothing of this scope in recent memory.”
Based on long experience, Jones said, this looks like a failure inside DoubleClick’s applications, not the result of hacking or malware.” However, he added that he sensed that Google was slow to respond, and may have only been alerted via Tweets about the problem.
Jones indicated that analysis by Dynatrace indicates that, generally, access to ad content “resulted in long delays followed by an ‘Error 502’ response, which was evidence the issue was within DoubleClick’s infrastructure and not with the World Wide Web.
“Dynatrace provides a proactive detection service,” he said. “We continually access sites and measure delays. This approach allows companies to detect problems early and switch to other ad-delivery companies.”
The process mating Web pages with advertising generally requires the Web page to load from the page owner’s site or from a forwarding agent such as Akamai Technologies. Ad content is added by third-parties such as DoubleClick. But with the outage, the content was no longer available.
This created a long time-out, which kept Web pages from opening all together, or resulted in blank Web pages as well as blank spaces where ads were supposed to be. But then the ad-servers started sending out error messages, which had the effect of speeding up page display, with no delays from the usual download of ads with rich media.
Recent cloud-level Web service failures have all been the result of human error. Typically a software update has been triggered with either incorrect parameters or inadequate testing. This may be yet another example of such problems.