Home > Networking

IDG News Service - The string of outages, malfunctions and bugs that have hobbled Twitter for the past week and a half prompted the company to issue another apology and explanation for its technical woes on Tuesday afternoon.

In an official blog post, Twitter identified key factors that have contributed to the problems, such as a spike in usage caused by the World Cup and the discovery of bugs while performing both short-term and long-term system enhancements and upgrades.

"As we go through this process, we have uncovered unexpected deeper issues and have even caused inadvertent downtime as a result of our attempts to make changes. Ultimately, the changes that we are making now will make Twitter much more reliable in the future. However, we certainly are not happy about the disruptions that we have faced and even caused this week and understand how they negatively impact our users," the post reads.

It's the second such "mea culpa" statement Twitter has issued in less than a week. On Friday, Twitter chalked up the problems experienced until then to faulty planning, monitoring and configuring its internal network -- issues it proceeded to address by doubling the network's capacity, fine-tuning its monitoring and re-calibrating its load balancing.

The problems, which started two weekends ago and continued Wednesday morning, have included complete outages of the Twitter site and its third-party application platform, intermittent availability issues, fixes and upgrades gone bad and bugs affecting specific features and capabilities, according to entries in the Twitter Status blog.

"I think it is disgraceful that Twitter is having these ongoing outage issues. Being transparent about their problems is nice but it is not a substitute for getting it right," said Mike Gualtieri, a Forrester Research analyst, via e-mail.

"My sense at this point is that there is a technical leadership and management issue that needs to be solved in order to solve the outage issues," he added.

However, one must keep in mind that the problems have been far from catastrophic and that Twitter is a free service, said Ken Godskind, chief strategy officer at AlertSite, a Web site performance monitoring company.

According to AlertSite's monitoring of Twitter's log-in process, the average response time in April was 7.15 seconds with a 98.6 percent success rate for loading the user's account page. In May, the numbers were 7.29 seconds and 99.68 percent availability. This month so far, the availability is down slightly to 98.46 percent, but the average response time is much faster: 5.92 seconds.

"In Twitter, I see a growing young company trying to keep up with demand and trying to build a sustainable business model," he said in a phone interview.

$(document).ready(function() { var numParas = $("#article_body,#insider_article_body").children("p").length; if (numParas <= 4) { $("#article_body,#insider_article_body").children("p:eq(0)").after($("#inset_module")); } else { $("#article_body,#insider_article_body").children("p:eq(3)").after($("#inset_module")); } });