Web 2.0 darling Twitter has long struggled with stability, a problem which came to a head when millions of users were faced with an error screen that lasted late into the morning today as the startup moved its system off Joyent’s hardware starting at 10 pm yesterday evening, in a migration effort that lasted until dawn.
“Although significant progress was made last night and today, a one-night push is not a miracle cure,” Twitter’s Biz Stone told me in email. He also reiterated Twitter’s goal of being a reliable tool, pointing to Lee Mighdoll’s hiring as VP of engineering earlier in the month and the fact that the new servers are hosted by Japan’s NTT.
Stone said that more information would be forthcoming and that the very small team of six are “really heads down” working on the issues. The move came as something of a surprise, said Joyent’s CEO David Young over the phone — especially considering the free capacity the company had been extending to Twitter — coupled with Twitter’s recent assertions of undying affection.
“We have many, many customers that are far bigger users of computing and bandwidth,” said Young. To blame the problem on Joyent’s hardware and services was, he said, “just silly.” Still, he reiterated his love of the service and said the Twitter team has “built something really special.”
But in all of the expressions of amicability, the subext of the posts remind me of publicist statements during celebrity break-ups. In response, Young joked, “If this is a celebrity breakup, then they’re the J. Lo and we’re one of the early husbands.”