What Happened During the ChatGPT Service Disruption

OpenAI confirmed on Tuesday that ChatGPT was experiencing disruptions across several parts of the platform. Users reported trouble accessing Codex, Custom GPTs, and workspace analytics tools, with the outage cutting across multiple features rather than a single isolated service.

Which FedRAMP Services Were Affected

According to OpenAI's status page, the company acknowledged ongoing issues specifically within FedRAMP workspaces. The affected services included Codex, workspace analytics, conversation search, the ability to search for Custom GPTs, ChatGPT user invitations, and the download endpoint for the Compliance Logs Platform.

Core Functionality Restored, But Problems Linger

OpenAI later reported that core functionality had returned to normal, even though certain issues remained unresolved. The company said it continued to see problems within FedRAMP workspaces affecting the same set of tools — Codex, workspace analytics, conversation search, Custom GPT search, user invitations, and Compliance Logs Platform downloads. OpenAI stated its teams were actively working on fixes and would share updates as more information became available.

User Reports Spiked Between 1 and 2 P.M.

Outage-tracking service Downdetector recorded a noticeable jump in user complaints, with reports climbing between 1 p.m. and 2 p.m. Roughly 75 percent of the reports centered on ChatGPT itself, while 12 percent pointed to issues with DALL-E and 8 percent involved the mobile app. Complaints ranged from interfaces that loaded slowly to users being unable to reach the platform at all.

A Problem That May Have Been Building for a Week

OpenAI's status page suggests the underlying issue had already been under investigation for about a week before Tuesday's broader outage became visible to users. This timeline indicates the FedRAMP-specific problems likely existed before the disruption spread to the wider user base. OpenAI's engineering teams said they were continuing to work toward a full resolution across every affected service.