Claude outage on May 19, 2026
Claude users likely saw failed connections, degraded availability, or session instability for this AI Tools service. The incident reached Minor severity and lasted 54 min. The stored report window does not show a preserved user-report spike.
Anthropic’s official status page is reporting a minor incident, “Elevated errors on Claude Haiku 4.5,” currently in monitoring, affecting claude.ai, Claude Console (platform.claude.com), Claude API (api.anthropic.com), Claude Code, and Claude Cowork. Downrightnow’s signals are otherwise quiet (0 user reports, DNS resolving normally, no major regional network anomaly, and no Hacker News chatter), so this may be intermittent or component-specific despite the potentially broad scope.
Cached · Tue, May 19, 2026, 08:03 AM UTC · scope: this incident
Status changed to Possible Issues
The stored readout moved to Possible Issues with confidence 38.00.
Incident detected
Minor confidence crossed the incident threshold for Claude.
Status changed to Operational
The stored readout moved to Operational with confidence 80.00.
Marked resolved
The incident was closed after 54 min of observed impact.
Peak confidence reached
Confidence peaked at 80.00 while the service was marked Operational.
Status changed to Possible Issues
The stored readout moved to Possible Issues with confidence 38.00.
No per-source signal payloads were preserved in this incident's snapshot window.
Confidence scoring
24.00The incident reached 24.00 peak confidence across 17 stored status snapshots.
User reports
0 / 15mNo elevated user-report volume was preserved in the incident window snapshots.
Status samples
10 affected10 snapshots in the window showed degraded, recovering, or major-outage state before recovery.
Recovery signal
ResolvedRecovery was recorded at Tue, May 19, 2026, 08:33 AM UTC.