OpenAI
ChatGPT and AI APIs · Official status page (opens in a new tab)
- 2Failures in OpenAI's systems
- 0Its services disrupted by failures elsewhere
- Dec 11, 2024latest Nov 11, 2026Documented since
- 1 of 2Own failures with root cause confirmed
Counts cover incidents in this archive only. They are not a complete outage history or a reliability measure.
Failures in OpenAI's systems
Incidents where the documented failure was in OpenAI's own systems. Durations are measured to the end points the sources give.
DateIncidentDurationStatus
- Nov 11, 2026OpenAI: Unable to start new Workspace Agent runsDuration: 9 h 42 minto: Provider marked resolvedResolvedUnknown · Root cause unknown
Stated cause
The provider has not stated a cause in its status updates.
- Dec 11, 2024ChatGPT, API and Sora outage: a new telemetry service overwhelms KubernetesDuration: 4 h 22 minto: All services fully recoveredResolvedChatGPTOpenAI APISoraConfiguration change · Root cause confirmed
Stated cause
A new telemetry service deployed at 3:12 PM PST had a configuration that made every node in each cluster run resource-intensive Kubernetes API operations. The API servers were overwhelmed, taking down the control plane in most large clusters; DNS-based service discovery depends on the control plane, so services could no longer find each other.
Showing 2 of 2 documented here