Voltar a todos os posts

OpenAI Resolves ChatGPT Search and API Latency Issues

2025-08-12Unknown2 minutos de leitura
OpenAI
ChatGPT
Service Outage

OpenAI has successfully resolved an issue that caused a partial outage for ChatGPT's search functionality on Monday, August 11, 2025. After several hours of work, the company confirmed at 10:02 PM that normal functionality had been fully restored for all users.

Timeline of the Service Disruption

The incident unfolded over approximately six hours, with the OpenAI team providing regular updates as they worked to resolve the problem. The following is a chronological breakdown of the event.

Initial Problem Identification

At 3:54 PM, OpenAI first acknowledged a developing issue, noting that users were experiencing an elevated number of errors across certain services. The team immediately began working to implement a mitigation strategy. Throughout the next hour, they continued to focus on deploying a fix.

Mitigation Efforts and Progress

By 4:58 PM, the team had implemented some initial improvements. However, the core problem persisted. An update at 5:29 PM clarified that while the availability of the search feature had improved, it was still suffering from high latency. Further work was required to bring performance back to normal. By 6:12 PM, the team confirmed that increased latency with the search feature remained the primary issue.

Pinpointing the Cause

A more detailed update at 7:13 PM provided specific insight into the problem's scope. The issue was primarily affecting API calls that use web search and any ChatGPT prompts that triggered a search query. This resulted in significant delays in generating responses for users and developers relying on these features.

Deploying the Final Fix

After continued work, a breakthrough occurred. At 9:25 PM, OpenAI announced that a definitive fix had been deployed. The team then entered a monitoring phase to ensure the solution was stable and effective. Following a successful monitoring period of just over half an hour, the issue was officially marked as resolved, concluding the service disruption.

Ler post original
ImaginePro newsletter

Assine nossa newsletter!

Assine nossa newsletter para receber as últimas notícias e designs.