Twitter was unavailable for users across the globe on Tuesday morning, with the site evidently suffering a total outage along with serious access issues lasting over an hour. Access to the service starting failing over the web, mobile and its API at 8:20am GMT, with error messages warning the network is both “over capacity” and is dealing with an “internal error”. By 10:00am, the majority of the service had returned to some sort of normality, with the company’s image handling service and home timelines still affected.
Twitters own status board updated at 9:00, identifying the outage, and Twitter’s developer-facing monitor confirmed that four out of the five API’s were out, suffering a “service disruption.” Around 8:50, the search API was updated to “performance issues”. Around 9 am, a second API was upgraded to “performance issues”, while a number of users were able to periodically access the site. Though the site’s status yo-yoed, and a number of services were still reported out more than two hours after the outage started.
First, the company confirmed the outage by, somehow, tweeting, from its support account. We were not able to view the tweet, because the site was down. In the early days, Twitter outages would happen quite often due to the company’s “over capacity” error message picked up a nickname: the fail whale. The service’s structure prevented the company from simply expanding capacity by adding servers to its back end, and in result it would collapse often under the weight of its users during big events.