[Outages-discussion] Facebook RFO - WSJ version

[Outages]

Cut and paste from WSJ,

Network went down three times Sept. 17-28, blocking outside developers from accessing Facebook data

Facebook Inc, says a rare string of outages last month were the unexpected result of planned software changes.

The social network went down three times between Sept. 17 and Sept. 28, preventing users from logging on or accessing mobile apps through Facebook for up to two hours at a time. Facebook says the incidents weren’t related.

“In the last several weeks, we’ve fallen a bit short on our availability,” Pedro Canahuati, Facebook’s director of production engineering and site reliability, said in an email to The Wall Street Journal. “Each unrelated issue began as a planned code change that had unintended consequences in our ability to respond to server requests.”

The string of outages is rare, according to Internet-analytics firm Dyn. The Sept. 28 outage appeared to be the most serious because it spread globally, says Dyn. The firm tried to reach Facebook from 200 locations world-wide, and found just a handful up and running during the outage.

The outages blocked outside developers from accessing Facebook data, meaning some users couldn’t log into mobile apps such as Tinder and Getaround with their Facebook accounts.

The Sept. 28 outage reduced that day’s revenue at GSN Games, a maker of online games like GSN Casino and Bingo Bash, by 5%, said Chief Marketing Officer Kevin Frisch. Others like mobile video app Pheebo were unable to sign up new users for about an hour.

The issues underscore the importance of Facebook as a source of traffic for publishers and app developers. Facebook understands the root causes of each case and is working to prevent similar issues in the future, Mr. Canahuati said.

A Facebook spokesman declined to elaborate. It is unclear if Facebook will provide a public explanation for the outages as it has sometimes done in years past.

Facebook also experienced less severe problems with its service last month. On four other days, Facebook had problems delivering a feature to developers that enables apps to automatically receive updates in Facebook data. Just before the first outage on Sept. 17, Facebook reported a problem with its payments platform.

The demands on Facebook’s infrastructure have grown in recent years as its user base expands. Nearly 1.5 billion people access Facebook at least once a month, while Facebook-owned apps WhatsApp, Instagram and Messenger are adding users at a rapid clip.

Facebook has been investing in data-center technologies that rely on software that can be updated quickly to respond to a spike in users. This includes networking equipment powered by Facebook-written software that funnels data among its computers and between Facebook’s computers and the Web.

Write to Deepa Seetharaman at Deepa.Seetharaman at wsj.com

regards,
/virendra 

> On Oct 5, 2015, at 3:43 PM, Jay Hennigan <jay at west.net> wrote:
> 
>> On 10/5/15 3:31 PM, Outages wrote:
>> http://www.wsj.com/articles/facebook-says-planned-software-changes-caused-outages-1444058402
> 
> Behind a paywall. Workaround is to search Google for the string "facebook-says-planned-software-changes-caused-outages". WSJ whitelists if referrer is Google search results.
> 
> --
> Jay Hennigan - CCIE #7880 - Network Engineering - jay at impulse.net
> Impulse Internet Service  -  http://www.impulse.net/
> Your local telephone and internet company - 805 884-6323 - WB6RDV
> _______________________________________________
> Outages-discussion mailing list
> Outages-discussion at outages.org
> https://puck.nether.net/mailman/listinfo/outages-discussion
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/outages-discussion/attachments/20151005/f0ddb0a6/attachment.html>