Those of us in software know how important testing is, but its importance is sometimes lost -- or simply underestimated -- by others. In 2013 we had several major Web outages that demonstrate to us just how important it is to keep your website running smoothly. The winners of the Internet race can be determined by fractions of a second-- here's a list of websites whose performance issues caused them to fall behind the pack.
The highest profile outage of the year (and possibly the decade) easily belongs to HealthCare.gov. From its launch on October 1st, the site was plagued by availability and performance problems, and given its direct connection with the hotly debated Affordable Care Act (ACA), not one hiccup went unnoticed. It literally took an act of Congress to iron out the site’s performance issues and ensure its future stability. While the site now appears to be operating for most, for the laws to be successful the website needs to operate as smoothly as possible — a point underscored by the fact that President Obama ends each progress meeting with the statement, “ . . . if the website doesn’t work, nothing else matters.”
Motorola -- Cyber Monday Outage
Just a few weeks ago, on Cyber Monday, shoppers hoping to score a heavily discounted Moto X smartphone were met with a non-functional purchase page. The page remained unavailable throughout the day forcing Dennis Woodside, Motorola’s Chief Executive, to apologize and explain that unanticipated customer demand and insufficient pre-sale site testing resulted in the outage. To compensate, Motorola promised to rerun the promotion twice while also doubling the number of phones made available — easily making them Cyber Monday 2013’s biggest loser.
Bank of America
On February 1st, Bank of America’s online banking services went down — for both desktop and mobile users. For a period of time the bank itself couldn’t even see users accounts, an issue compounded by the fact it was a Friday and the first of the month — preventing customers from accessing their paychecks in time for the weekend. While their online portal was back up a few hours later, the damage was already done via social networks — forcing Bank of America to not only apologize for the outage but also assuage fears that the outage was a result of malicious hackers — which it wasn’t. Nonetheless, the outage holds up as one of the most disruptive of 2013.
Earlier this month, on December 10th, Yahoo mail went down for a large number of users, remaining unavailable for many through December 12th. This followed an October website redesign that had already left many long time account holders feeling less than impressed. In the ongoing battle for Web Mail supremacy, performance is sure to play a major role in attracting and retaining users — it’s possible that this latest outage cost Yahoo significant ground in the race to the top.
Amazon Web Services
Friday the 13th turned out to be an especially unlucky day for Amazon Web Services. This was the second major outage of 2013 for the cloud infrastructure provider, whose clients include many start-ups but also major enterprises and government agencies the likes of NetFlix and NASA. While Amazon was able to restore service in just under 2 hours, it served as another reminder of the inherit risks of relying exclusively on cloud infrastructure.
The August 14th Outages
The August 14th outages are really 3 separate failures that affected Google, Amazon and Microsoft on the same day. Microsoft’s outage continued on for 3 days, forcing users to put up with numerous intermittent Outlook.com and Exchange ActiveSync issues. Amazon.com went down for roughly 25 minutes — costing the e-commerce giant approximately $1,104 of revenue per second. Google’s outage, while lasting just a touch over 1 minute, resulted in a 40% drop in global Internet traffic over its 61-second outage. You can read Scott Barber’s take on the outages, and their possible causes, in his SmartBear blog post from August.
At about noon Eastern Time on Thursday August 22nd, the Nasdaq went dark — an unprecedented outage that lasted nearly four hours. Not only was Nasdaq trading halted during the outage, but all NYSE trading was suspended as well, to reduce the risk of market repercussions. While there was speculation that the outage was a result of malicious hackers or a market-wide crash, in the end, the cause was a data feed failure. This is clearly an event that all financial companies should contemplate when implementing monitoring and testing tools.
Just a couple days after Facebook’s SNAFU, LinkedIn made similar headlines. At approximately 11am on October 23rd LinkedIn users started to experience performance and availability problems that are rare for the site — preventing them from logging into their accounts and causing the site to load incorrectly. Roughly 4 hours later LinkedIn had corrected the problem — plenty of time for the Twitter-verse to have some fun at the social site’s expense.
Starting at around 10am on Monday, October 21st, Facebook users were unable to update their statuses, like posts, send messages or comment on items in their timelines. Over the course of the 2-hour performance interruption many users took to Twitter to vent their frustrations and comment on life without Facebook.
At approximately noon on November 12th, LivingSocial.com went down and remained unavailable for close to two days — preventing customers from purchasing the various deals and discounts they offer. The outage was another setback on the road to profitability, which had been derailed earlier in the year via an April cyber attack that forced 50 million users to reset their passwords — costing millions in revenue. Despite the length and business impact of the outage, LivingSocial’s site issues flew largely under the radar.
What are your thoughts on some of these major website outages? Were you personally inconvenienced by any? Let us know what you think in the comments.