Claude Outage History

An outage disrupted Claude's Admin API and the Console, with usage reporting and cost reporting affected. Users attempting to manage accounts, access reporting screens, or run queries against the Sonnet 4.6 model reported errors and slow responses. StatusGator logged an Early Warning Signal starting at Mar 3, 2026 3:44 PM UTC, roughly 1 hour and 9 minutes before the first official acknowledgement at 4:53 PM UTC.

Engineering teams pushed fixes over the next hours. The admin API was operational again by 12:58 AM UTC and the Console (platform.claude.com) followed by 1:34 AM UTC. A separate update addressed usage reporting, and by 8:33 AM UTC the service had fully recovered. The outage affected Claude API (api.anthropic.com) and platform.claude.com, with cost reporting also impacted during the incident.

The Claude service experienced an outage starting around 2:53 PM UTC with elevated errors on Claude Haiku 4.5. By 3:43 PM UTC the issue expanded to Claude Sonnet 4.6 and related components, including Claude API (api.anthropic.com), claude.ai, and Claude Code, with users seeing failures and degraded performance. The incident affected users globally, who reported a range of symptoms from timeouts, failed uploads, and unresponsiveness across regions such as the United States, United Kingdom, India, United Arab Emirates, and the Philippines. The status updates described elevated errors and degraded service during this period, and recovery began by 4:43 PM UTC as components returned to normal.

User reports surged in the late afternoon UTC, with activity peaking around 4:00 PM UTC to 4:15 PM UTC, spanning regions from Dubai to the United States, United Kingdom, India, and the Philippines. Many users described errors such as inability to run scripts, network timeouts, and unresponsive AI responses, indicating widespread impact before stabilization.

Users across multiple regions encountered degraded Claude services as elevated errors affected Claude Opus 4.6 components, including claude.ai, platform.claude.com, Claude Code, and the Claude API (api.anthropic.com). The impact manifested as tools not loading, chats failing to load or generate content, and authentication or task-generation issues. Beginning around 3:18 AM UTC, reports flowed in and spread through the morning, indicating a broad, global disruption.

Status updates from Claude showed an ongoing investigation with fixes being implemented and monitored. A series of remediation steps began to restore functionality, and by 2:33 PM UTC most affected components were returning to normal operation (claude.ai, platform.claude.com, Claude Code, and Claude API). The incident represented a partial outage across Claude Opus 4.6 services rather than a full service shutdown.

Users experienced degraded Claude performance across the Opus 4.6 stack, including the Claude API, claude.ai, platform.claude.com, and Claude Code. The outage led to incomplete or failing response generation, with reports of chats not loading, uploads failing, and older chats disappearing in some cases. The impact appeared globally across multiple user locations and activities, illustrating the breadth of the disruption.

An outage began around 11:37 AM UTC with Claude reporting service unavailability due to server errors. Users worldwide could not access Claude.ai, and login/logout flows along with the chat experience were affected. The disruption extended to Claude Console and Claude Code, while the Claude API was noted as functioning for some operations, indicating the issue was concentrated on user-facing paths rather than the API itself. Status updates emphasized elevated errors across multiple Claude components.

On Feb 28, 2026 6:13 PM UTC, Claude Opus 4.6 reported elevated errors affecting Claude API at api.anthropic.com, claude.ai, and Claude Code. Users interacting with Claude features may have experienced errors or degraded functionality during this period.

The issue prompted an investigation, and the affected components returned to normal by 6:38 PM UTC, approximately twenty-five minutes after the event began. The incident appeared localized to Claude Opus 4.6 components rather than indicating a broader service outage.

Users experienced degraded access to Claude.ai beginning around Feb 28, 2026 12:58 PM UTC as elevated errors affected claude.ai and some users could not load the Web/API endpoints or the settings page. Status updates showed a partial outage across the Claude.ai component, with reports of connection timeouts and upstream errors from multiple regions. StatusGator issued an Early Warning Signal at Feb 28, 2026 12:52 PM UTC tied to settings page connection problems (ending Feb 28, 2026 1:03 PM UTC), providing about six minutes of lead time before the first official acknowledgment at Feb 28, 2026 12:58 PM UTC.

Claude users across multiple services encountered an elevated error rate that caused failed requests and degraded performance for a short period. The Claude status page showed the issue beginning at 7:33 PM UTC when the affected components Claude for Government, Claude API (api.anthropic.com), claude.ai, platform.claude.com, and Claude Code were flagged as experiencing elevated errors and the page indicated an ongoing investigation. The issue was resolved about ten minutes later, with status returning to normal by 7:43 PM UTC. User reports from multiple regions supported the window of impact, including messages around 7:34 PM UTC from Inhambupe, Bahia, Brazil and 7:42 PM UTC from Surat, Gujarat, India.

The Claude platform experienced a coordinated outage affecting usage reporting across Claude API (api.anthropic.com), Console (platform.claude.com), Claude AI (claude.ai), and Claude Code. Users encountered missing usage data and degraded interactions across dashboards and APIs. The incident began on 3:58 PM UTC and persisted until 5:18 PM UTC, spanning roughly 25 hours and 20 minutes. The provider issued 16 status updates during the incident as engineers investigated and rolled out fixes across the affected components.

A sequence of fixes gradually restored functionality, including a targeted fix for the /v1/organizations/usage_report/messages admin API endpoint and subsequent hotfixes for Claude Code on Windows. Elevated errors on claude.ai were addressed, and by 5:18 PM UTC all services returned to normal operation.

Claude experienced a prolonged outage affecting multiple components including claude.ai, platform.claude.com, Claude Code, Claude API (api.anthropic.com), Claude for Government, and the Claude Desktop app. Users faced widespread unavailability, crashes, and API errors across chat, coding, and government workflows. The outage began on 1:58 PM UTC and was resolved by 11:58 PM UTC.