ChatGPT down for 90 minutes due to a DDos attack

OpenAI attributed the recent widespread access failure of ChatGPT to an external cyber-attack. As reported by foreign media outlets including CNBC in the U.S. on Thursday (local time), OpenAI disclosed the detection of indications pointing to a distributed denial of service (DDoS) attack. This type of attack causes overload by initiating simultaneous access to the platform from external sources.

OpenAI stated, “We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack.” Importantly, OpenAI clarified that user information remained unaffected. ChatGPT boasts a weekly active user base of around 100 million, with over 92% of Fortune 500 companies utilizing the platform.

The ChatGPT connection failure persisted for approximately 90 minutes around 9 a.m. Eastern time on Wednesday. Users attempting to access the website encountered a message saying, ‘ChatGPT is at capacity right now.’ This incident occurred two days after OpenAI hosted a developer conference in San Francisco, where they unveiled their enhanced model, GPT-4 Turbo. The impact of this failure extended beyond ChatGPT users, affecting the 2 million developers who build their tools using OpenAI’s application programming interface (API).

Approximately 90 minutes after the failure, OpenAI declared that they had identified and resolved the issue, ensuring the service was operating normally. However, it was reported that some users continued to encounter error messages after the announcement.


whatsup@donga.com