Twitter changes cause widespread outages after Elon Musk takeover


SAN FRANICSCO — Elon Musk’s Twitter is a house of cards.

Twice recently, almost exactly a month apart, minor changes to Twitter’s code seemed to break the website.

The latest outage came on Monday as thousands of users found they couldn’t access links, photos or other key aspects of the site.

“One small API change had massive ramifications,” said Twitter CEO Elon Musk. writing in a tweet on Monday, referring to the tool used by third-party developers who run programs that take data from Twitter and post it to its site. “The code stack is extremely fragile for no good reason. Will eventually need a complete rewrite.

It was the second time on Monday that he had turned to this explanation, both times call the “fragile” site.

Since taking over Twitter, CEO Elon Musk has laid off more than two-thirds of the company’s staff, embarking on aggressive cost-cutting and laying off workers in part by forcing them to commit to a workplace. work “extremely hard” or leave the company. The massive layoffs have sparked widespread concerns about Twitter’s ability to retain core functions as critical engineering teams have been reduced to one or zero employees.

In the months following the takeover – and subsequent layoffs – Twitter faced multiple outages, hampering key functionality: loading tweets and notifications, sending tweets and direct messages, accessing links and pictures. Each has been told — by current and former employees, or by Musk himself — to come as the company makes changes to its code.

“Every error in code and operations is now deadly,” a former engineer told The Washington Post in November, explaining that those who remained “were going to be overwhelmed, overworked, and because of that, more likely to make mistakes.” . The former engineer spoke on condition of anonymity for fear of reprisals.

Prior to the Musk takeover, the company had a risk assessment team that reviewed product changes for anticipated issues. Twitter’s risk assessment process was intended to flag potential issues before they arise. But the team was fired after Musk’s takeover, the Washington Post reported, resulting in bug-riddled product rollouts.

Musk and Twitter did not immediately respond to requests for comment.

Since taking over Twitter, Musk has implemented a plan to cut the company’s staff by 75%, aggressively cut costs and pursue new revenue streams, such as charging $8 a month for business blue verification icons. But his tenure has also been marred by embarrassing incidents, such as the botched rollout of the checkmark feature, which resulted in a swarm of copycats and prompted Twitter to temporarily suspend the subscription service on several occasions.

Musk sued Twitter by pledging to restore ‘free speech’ to the platform, firing the company’s previous list of officials he accused of a content moderation approach rooted in protection against the harms of hate speech and disinformation. Musk has also promised transparency on the old regime’s decisions – such as the effort to limit the circulation of a New York Post article about the contents of Hunter Biden’s laptop – but has repressed about company information leaks under his own direction.

Even before Musk’s takeover, Twitter employees had warned of the site’s vulnerabilities in the event of an outage. Twitter whistleblower Peiter Zatko warned of a potential occurrence of overlapping outages at Twitter’s offsite data centers, for example, in a complaint obtained by congressional committees.

This type of outage, he said, could leave critical data unrecoverable — and cause Twitter to go down for months. Despite concerns about Twitter’s vulnerable infrastructure, Musk ordered the shutdown of Twitter’s largest data center in Sacramento in December, The Washington Post reported at the time.

On Monday, Twitter users were greeted with problems almost as soon as they opened the site.

When users clicked on a link on Twitter, they received the message: “Your current API plan does not include access to this endpoint”, it read and directed them to a developer page.

On the Down Detector website, which tracks online outages, complaints poured in: “User reports point to problems on Twitter,” he said.

“We made an internal change that had unintended consequences,” Twitter’s support account said in a tweet.

By late morning, some of the functions seemed to have been restored.

“Things should now be working normally,” the Twitter support account said in a tweet. “Thank you for staying with us!”

Monday’s pattern reflected an early February outage, which came as Twitter made similar changes to its Developer Data Feed API.

Twitter faced a widespread outage on Feb. 8 that left users unable to tweet and direct message, follow other accounts, and upload content to their feeds.

“As of February 9, we will no longer support free access to the Twitter API, both v2 and v1.1. A basic paid tier will be available instead,” Twitter writing this month.

Musk said Twitter was trying to crack down because Twitter’s freely available data was “abusedby bots peddling scams, but later said the company would make a free version available, as criticism poured in on its aggressive efforts to monetize aspects of the site that were previously free.

This set of issues follows a widespread global outage that Twitter faced in December.

In group discussions among current and former engineers at the time, some speculated that the December outage happened after a software update went wrong.

Monday wasn’t the first time Musk suggested that Twitter’s code needed to be completely rewritten. He has maintained that position for months, since taking over the site last year. During a December Twitter Spaces the site’s live audio feature, he said the company’s code base needed an overhaul.

Pressed by an attendee to explain what that meant, Musk got irritated.

“Amazing, wow,” he said, after hesitations and pauses. “You are a moron. … What a moron.”

Leave a Comment