The Wayback Machine - https://web.archive.org/web/20101229114042/http://royal.pingdom.com:80/2007/12/19/twitter-growing-pains-cause-lots-of-downtime-in-2007/

Pingdom Home

US + international: +1-212-796-6890

SE + international: +46-21-480-0920

Business hours 3 am-11:30 am EST (Mon-Fri).

Royal Pingdom

Twitter growing pains cause lots of downtime in 2007

Twitter logoIn most ways, 2007 has been a great year for Twitter. The service has grown tremendously and has become one of the big social network successes.

The flip side of this popularity is that they have often been unable to handle all the traffic their large user base has been bringing in, which has affected the stability of their website. In other words: Twitter has been suffering from growing pains.

We have monitored the availability of the Twitter.com website since mid February 2007 until today (and counting). So far, the accumulated downtime of the website totals 5 days and 23 hours. This includes maintenance, but is still a very high number. So far in December, the website has been unavailable for a total of almost 11 hours (part of which was due to problems related to a data center move).

The worst month for Twitter by far was March, when the Twitter.com website totaled over 2 days and 10 hours of downtime. This is also the month when Twitter’s popularity really took off, winning the SXSW Interactive Festival, so it looks like the rapid growth that month took its toll.

Twitter.com downtime in 2007
Month Uptime (%) Downtime (d, h, m)
December 2007 97.51% 10 hours, 59 mins
November 2007 98.74% 9 hours, 4 mins
October 2007 99.02% 7 hours, 19 mins
September 2007 99.06% 6 hours, 44 mins
August 2007 98.47% 11 hours, 22 mins
July 2007 99.59% 3 hours, 2 mins
June 2007 99.08% 6 hours, 37 mins
May 2007 97.67% 17 hours, 9 mins
April 2007 99.11% 6 hours, 15 mins
March 2007 92.09% 2 day, 10 hours, 3 mins
February 2007 98.21% 6 hours, 22 mins
TOTAL 98.06% 5 days, 23 hours, 0 mins

Above: Note that the February data is not complete, and that the December data of course isn’t complete as of this writing. This data can be found, always up to date, on the public uptime report page for Twitter.com. (A day-by-day breakdown can be had by clicking on the link for a month.)

Another indication that the Twitter.com website has been suffering from more traffic than it can handle is that a lot of days the website has been unavailable for numerous short periods of time, often ranging from 1 to 10 minutes, spread over large parts of the day, which indicates that the website has been struggling to keep up with demand and deliver pages during much of that time.

This last weekend Twitter moved their service to a new data center, which will hopefully alleviate the problems they have been experiencing.

Will the Twitter website be more reliable from now on? As we mentioned, you can keep track of this on a public report page from Pingdom that shows the uptime of the Twitter.com website and is updated continuously.

Note: Twitter’s recent downtime and data center move has received a lot of buzz and speculation, for example in Mashable, Ensight and Data Center Knowledge.

Want to test your site every minute?








You will get an email with your login information.

23 Comments

December 20, 2007 3:11 pm

Twitter Downtime On the Upswing

January 15, 2008 1:08 pm

Steve Jobs Speaks. Twitter Goes Down.

May 24, 2008 12:59 pm

The Mind of Dave Winer

Why email won’t die anytime soon

EmailIf you follow the tech media, you’ll know that every few months, some journalist or blogger will start speculating about the imminent demise of email. Headlines along the lines of “Email is Dying” or “The Death of Email” show up in RSS feeds all over the place. You know the drill. This has been going on for years and we’re surprised this argument hasn’t (pardon the pun) died out by now.

Email is most definitely not dying, and here’s why.

Read more

The best Royal Pingdom posts of 2010 (Happy Holidays!)

Happy Holidays from Pingdom

Happy Holidays everyone!

Like much of the world, we here at Pingdom will be taking a short Christmas break. Since we won’t be updating the blog until sometime next week, here is a selection of posts from the past year that you might have missed.

Read more

The major incidents on the Internet in 2010

Internet Incidents

In what has become something of a yearly tradition, it’s now time for us to present 10 of the most noteworthy incidents on the Internet from this past year. As you’ll see, 2010 has been very interesting.

Just like previous years, we have included problems ranging from website outages and service issues to large-scale network interruptions. If you’re an avid Web user, you are bound to recognize several of them.

Read more

The most reliable (and unreliable) blogging services on the Web

Blogging services

Blogging services have been around for a long time, with pioneers like Blogger paving the way for Wordpress.com and more recent arrivals like Tumblr and Posterous. There are millions upon millions of blogs out there, many of them residing on these services.

One big bonus of using a blogging service is that they take much of the pain away from having a blog since they handle the hosting for its users and everything is already set up. Once you publish, the responsibility for keeping that content available online rests firmly on the shoulders of the blogging service.

With that in mind, we decided to test five of today’s most popular blogging services to see how reliable they actually are.

Read more

The most popular web servers for REST APIs

Web server softwareRESTful APIs have become increasingly popular both among web services and developers and are easy to serve up with the same software used for regular web pages. In May of 2010, 74% of web APIs used REST as their protocol.

When setting up servers for a REST API it can make sense to use a web server software that is a bit more lightweight than what you’d use for a full-blown website. The gains are, at least in theory, that each API server that way could handle more requests since it would be less taxing on system resources.

But is that what actually happens, or do most web services just put up an Apache server, same as they would do for a regular website?

Read more