Claude Outage History

Claude.ai experienced a minor issue where responses appeared to "hang" for approximately 5 seconds after text finished streaming before messages completed, starting around 20:00 UTC on March 20. The issue only affected the Claude.ai web interface and did not impact the Anthropic API. The incident was resolved after 2 hours when the identified fix was successfully implemented.

Claude experienced elevated error rates affecting both Opus and Sonnet 4.6 models for 1.5 hours, impacting claude.ai, the developer console, and API services. The team identified the underlying issue and gradually restored service, with Opus 4.6 returning to normal first followed by full resolution of errors across both models.

Claude Opus 4.6 experienced elevated errors affecting both claude.ai and api.anthropic.com for 19 minutes. The incident began at 8:53 PT and was resolved by 9:14 PT, with error rates returning to normal baseline levels.

Claude Sonnet 4.6 experienced elevated errors for 4 minutes starting at 8:23 PT / 15:23 UTC on March 19, 2026. The incident affected both claude.ai and api.anthropic.com services. The error rate returned to baseline levels by 8:42 PT / 15:42 UTC, resolving the issue.

Claude experienced elevated authentication errors from 23:59-00:30 UTC, preventing users from logging into claude.ai, the console, API, and Claude Code, with additional issues including image upload failures and unresponsive interfaces. The incident affected all major Claude surfaces for 1.1 hours before being resolved, with the team continuing to monitor for further issues.

Claude experienced elevated errors affecting claude.ai, the developer console, and API services for 6 minutes on March 19th. Users reported inability to sign in, upload documents, create prompts, and access their paid accounts across multiple browsers. The incident was investigated and resolved within the 6-minute duration.

Claude.ai experienced a major service outage lasting 49 minutes, causing elevated errors that prevented users from accessing the service with authorization/oauth failures and "try again later" messages. The incident also affected Claude Code login/logout functionality during the outage. A fix was implemented and monitored, leading to full resolution of the service disruption.

Claude experienced increased errors on Opus 4.6 that affected claude.ai, console.anthropic.com, and api.anthropic.com for 1.5 hours. Users reported that Claude APIs were not working during this time. The issue was resolved after implementing a fix and monitoring the results.

Claude Opus 4.6 experienced elevated errors for 3.1 hours, affecting claude.ai, console.anthropic.com, and api.anthropic.com. The service team investigated the issue and implemented a fix, with monitoring ongoing to ensure resolution.

Claude Opus 4.6 experienced elevated errors for 4.7 hours, affecting claude.ai, console.anthropic.com, and api.anthropic.com. The issue was identified and mitigations were employed while investigating the root cause, with a fix ultimately implemented and monitored until resolution.