The Case of the Fail Whale

I bet that on Sunday afternoon, between 1:30 PM and 4:30 PM EST, Twitter will either be down for part of the time or third party applications like Tweetdeck and Seesmic Desktop will only update sporadically. Why am I making this bet, it is because the World Cup final will being occurring then . Ever since the World Cup started Twitter has had trouble staying up. This is disappointing, but not surprising. It has happen again and again, when ever there is a big event. Their solution to this problem was to drop the API call, they dropped it from 350 calls to 175 and at sometimes it was as low as 75 calls per hour. The bottom line for the end user is that those third party application they depend become unusable, with the message api limit has been met constantly popping up. Twitter says it is limiting the api call to help all Twitter user. Unfortunately limiting API calls hurts the power user the most. These are the people that have thousands of followers and depend on third party applications the most. For them the thought of using the Twitter Web site is laughable. Not only does api limiting hurt basic desktop applications, it also hurts those applications that people use to get information off such as Mr Tweet and Socialtoo.

Two years ago, when Twitter was new and used mainly by the first adopter crowd, this sort of behavior was accepted as part of their growing pains. However two years later, Twitter has gone main stream and it is still having many of the same problems, (albeit less often)this is no longer acceptable. Neither is their solution of throttling the api call every time there is a new big event. Especially when the big event was something they could have planned for, such as the World Cup. The World Cup is the most popular sporting event in the world outside the Olympics, anyone outside the U.S. knows this. Twitter should have planned for a large increase in tweets as soon as the World Cup was scheduled. Twitter long ago stopped being a application used by early adopters and geeks. It is now part of the main stream and needs to start acting like it. They either need to purchase more servers. If that is not practical, then they need to set up arrangement with multiple server companies to be able to rent servers when necessary. What ever their solution is , it can’t be what they do now. Their present solution just leaves the end user frustrated and angry.

3 thoughts on “The Case of the Fail Whale

  1. This definitely sounds like Geek speak what the hell is API? and you say that it is laughable that people actually use the Twitter web site?? Must look up Mr Tweet and Socialtoo.

  2. I stated at the start of the World Cup that Twitter need be operated like a national power grid. When a major event happens they need a system of routing in more power and bandwidth. The UK power grid will be doing this for the World Cup final, especially for half time and full time to cope with all the cups of tea being made. If they can manage it, surely Twitter can.

  3. My understanding is that this issue is not a server/bandwidth issue but a software architecture issue and can’t be solved just by adding more throughput or even more CPU at it.

Comments are closed.