As performance continues to be sluggish on Twitter, the platform suffered from a major outage last night.
Elon Musk’s reign as CEO of Twitter has been nothing short of catastrophic for the platform. Shortly after he bought the platform for $44 billion, Musk laid off huge swaths of his staff. Ever since then, the performance of the platform has been sluggish. Pages have been slow to load, media fails to load, and feeds sometimes kick out error messages. Rather than focus on fixing what ails the platform, Musk’s top priority seemingly continued to be the continued torching of the platform. One recent example was the mass banning of journalists.
The biggest beneficiary of Musk screwing up everything has been Mastodon. Currently, the platform is on the verge of surpassing 9 million users. As of this writing, the current user count is estimated to be at 8,922,572 accounts:
Musk, for his part, in the words of some critics, has simply continued to chew through the Twitter wires. At one point, he even boasted that he had pulled the plug of a critical server rack and the website continued to operate. For some observers, the question was how much longer could Twitter stay functioning before Musk finally finds the last straw in the system, pulls it, and pull the whole site offline. For critics, they got their answer late last night as Twitter suffered from a major outage.
According to DownDetector, at around 3:43PM yesterday, reports started coming in that the site went down. The downtime reports peaked at 4:43PM with 1,757 reports that the site was down in many regions around the world. The event went on until 8:28PM where the reports slowed down to the normal sluggishness. In a world where even a half an hour of downtime for a major website ends up making headline news, an outage that lasted nearly 4 hours is almost unheard of for a major website unless it was a massive hack.
Following the outage, Musk posted a tweet saying that they rolled out major back end architecture changes and that Twitter should feel much faster now.
Few are, of course, believing that line of reasoning. If you are a smaller website, chances are, you have done your research on a major upgrade, learned all of the ins and outs of how to integrate such a change, put in place some contingency plans in the event things go wrong, inform users ahead of time that this is going to happen and when, and carry through. Usually, a change might take about an hour to complete. If you have some customized caching system in place, it might take a little longer for the changes to take effect, then monitor for performance issues after.
For a larger web service, you more than likely have an offline version of the website. So, you make several tests to ensure that the transition from one system to another will be smooth. After that, the changes are gradually rolled out across the network of servers to make the transition from one system to another seem seamless and not noticeable to the users (at least as much as possible). Twitter is not a small website, so the idea that the whole site goes offline over some kind of change in the “architecture” seems unlikely. The last thing you want to do is test something in production in an environment like that, after all.
It appears that Musk and his team (whatever team he has left these days, that is) was able to bring the website back online again for now. For many observers, the question is, what’s the next stupid thing he does?
Drew Wilson on Twitter: @icecube85 and Facebook.
