On June 4th, a major outage affected the widely used chatbot ChatGPT for several hours.
This unscheduled downtime impacted millions of users worldwide who rely on ChatGPT for a variety of tasks, causing disruption to their workflows.
Developed by OpenAI, ChatGPT has garnered a massive user base exceeding 100 million. Users leverage ChatGPT for various tasks, including coding assistance, brainstorming ideas, and composing reports.
The outage, which began around 2.30 AM ET (7.30 AM WAT), affected users globally. Reports on Downdetector, a website tracking online service outages, revealed a surge in user complaints, with peak numbers reaching around 2,300 in the US and 1,000 in the UK.
Users attempting to access ChatGPT on the web or mobile app encountered error messages like “bad gateway” and “internal server error,” effectively rendering the service unusable.
OpenAI acknowledged the outage on its service status page. Initially, the status simply indicated that they were “investigating the issue.” Over three anxious hours later, the update changed to “continuing to work on a fix,” offering little solace to users desperately awaiting a resolution.
Finally, after nearly six hours of disruption, OpenAI provided a glimmer of hope by announcing that they had “identified the issue and mitigating.”
As of now, OpenAI has not disclosed the root cause of the outage. While talks swirled online, it appears the issue was not a simple server malfunction. Some users theorized a difference from a similar outage in November 2023, which was attributed to a distributed denial-of-service (DDoS) attack.
The outage greatly impacted ChatGPT’s user base, many of whom have become heavily reliant on the tool for various aspects of their work. This downtime likely provided a temporary boost to competitor chatbots like Google Bard and Microsoft Copilot, as users sought alternative solutions during the outage.
At the time of this report, ChatGPT service appears to be restored, but not yet stable. However, a clear explanation from OpenAI regarding the cause of the outage has not been given.
Author
