OpenAI: Degraded Performance affecting Agents API
The Agents API experienced degraded performance affecting managed sessions, which has been resolved. The issue was not related to a specific version or configuration.
– UTC
Here's why. Incidents from Monday, September 14, 2026 (UTC), ranked byBreakage Velocity.What is Breakage Velocity?A live score of how fast an outage or regression is spreading. The more people hit it, and the faster, the higher it climbs; it cools off as things calm down.Open-ended scale: under 25 low · 25+ elevated · 100+ high · 300+ extreme.
6 outages, 1 still active today. Worst: Anthropic (<1 min).
The Agents API experienced degraded performance affecting managed sessions, which has been resolved. The issue was not related to a specific version or configuration.
– UTC
A recent Windows update (September 8) caused Claude Cowork on Windows to be unable to run local commands, affecting its ability to reach the user's computer drive for workspace access. Most users can still use chat and file reading/editing functionalities, but no in-app workaround is available yet. Microsoft has developed a fix that will be released soon.
– UTC
Increased wait times for macOS jobs due to network errors and latencies. Mitigation implemented by the hosting provider has reduced wait times back to normal.
– UTC
Elevated error rates were observed for OpenAI's Codex and ChatGPT Work services, affecting all users. The issue has been resolved.
– UTC
Quay.io, a container registry service, has become read-only due to an outage affecting all users. No specific affected versions or workaround are provided in the report.
since
The Docker Hub Registry experienced an error (ESD) that prevented users from viewing newly pushed images and tags. The issue has been resolved, and all data is now accessible.
– UTC