AWS wellarchitected documentation change
Summary
Extended cost anomaly detection with custom metrics for generative AI workloads and alert routing
Security assessment
Changes enhance cost monitoring capabilities by adding AI-specific metrics like token counts. While this improves cost visibility, there's no evidence of security vulnerability remediation or security feature documentation. The focus remains strictly on financial observability.
Diff
diff --git a/wellarchitected/latest/financial-services-industry-lens/fsicost05.md b/wellarchitected/latest/financial-services-industry-lens/fsicost05.md index 08bf749ae..0ddb0b560 100644 --- a//wellarchitected/latest/financial-services-industry-lens/fsicost05.md +++ b//wellarchitected/latest/financial-services-industry-lens/fsicost05.md @@ -14,0 +15,2 @@ Anomalies can drive up cost. Set up AWS Cost Anomaly Detection to detect and ale +Extend Cost Anomaly Detection with custom metrics such as `token_in`, `token_out`, and `embedding_ops`, and route alerts to product or data owners when cost spikes correspond to new prompt deployments, unexpected retrieval-augmented generation (RAG) expansion, or fine-tuning jobs running out of schedule. Combine these alerts with CloudWatch dashboards to correlate generative AI usage trends with cost anomalies in near real-time. +