OpenAI incident: Degraded performance in ChatGPT, Codex, and APIs

1. resolvedOct 21, 2026, 03:28 UTC

   All impacted services have now fully recovered.

2. monitoringOct 21, 2026, 02:48 UTC

   We have applied the mitigation and are monitoring the recovery.

3. investigatingOct 21, 2026, 02:26 UTC

   We are investigating the issue for the listed services.

What APIStatus.watch saw

• Oct 21, 2026, 02:30 UTCoperational → degraded
• Oct 21, 2026, 02:50 UTCdegraded → operational

Other services with incidents at the same time

Overlapping incidents aren't necessarily related.

• Notion: Unable to install connections created by other workspace membersMajor · Oct 20, 2026, 18:46 UTC
• Elasticsearch: Elastic Cloud scaling issues in AzureMajor · Oct 20, 2026, 20:07 UTC
• Plaid: Service Disruption Impacting Chase connectionsMajor · Oct 21, 2026, 03:18 UTC
• Shopify: Some merchants encountering error messages when trying to login to their adminsMinor · Oct 8, 2026, 14:49 UTC
• Twilio: SMS Delivery Delays and Failures from a Subset of Twilio Alphanumeric Sender IDs to Mobile ChinaMinor · Oct 9, 2026, 06:42 UTC
• Twilio: Voice Call Post Dial Delay from a Subset of Twilio Phone Numbers to ThailandMinor · Oct 12, 2026, 11:53 UTC
• Twilio: Voice Call Failures and Post Dial Delay from a Subset of Twilio Mobile Numbers to BrazilMinor · Oct 19, 2026, 13:29 UTC
• Twilio: SMS Delivery Delays and Failures Between a Subset of Twilio Long Codes and Multiple Networks in ItalyMinor · Oct 19, 2026, 16:53 UTC
• Mapbox: Intermittent Routing Disruptions in US-EastMinor · Oct 20, 2026, 18:43 UTC
• Twilio: SMS Delivery Delays from Twilio to T-Mobile Czech RepublicMinor · Oct 21, 2026, 01:10 UTC

How often does OpenAI have incidents?

OpenAI reported 48 incidents in the last 90 days, 9 of them major or critical. A typical incident lasted 1h 34m. See OpenAI's uptime and incident history