2025-09-10: Error rate SLO violation for anthropic model inferences in ai-gateway

Error rate SLO violation for anthropic model inferences in ai-gateway (Severity 3 (Medium))

Problem: An outage with Anthropic caused a substantial spike in error rates for AI model inference requests via ai-gateway.

Impact: Error rates for AI model inferences through ai-gateway reached 17.66%, exceeding the service level objective and affecting users relying on these model outputs.

Causes: An outage with Anthropic's services led to increased errors for AI model inference requests.

Response strategy: The alert has resolved after Anthropic's services recovered.

---

This ticket was created to track INC-3846, by incident.io 🔥