Basic information
| Field | Details |
|---|---|
| English / full name | AWS Cost Anomaly Detection |
| Chinese | AWS 成本异常检测 |
| Japanese | AWS コスト異常検知 |
| Review priority | 复习优先级 ⭐⭐⭐ |
| Often confused with | AWS Budgets / Cost Explorer / CloudWatch |
In one sentence
Uses machine learning to learn normal spending patterns and identify significant deviations with cost impact and root dimensions.
Core summary
A cost monitor defines scope, the model creates anomalies, an alert subscription applies notification thresholds, and people investigate and respond.
Study points
- Scope monitors by service, member account, tag, or cost category.
- Individual alerts use SNS; daily and weekly summaries use email.
- Processed cost data can introduce about a 24-hour delay.
- Use Cost Explorer, CloudWatch, CloudTrail, and change records to find the actual cause.
Common pitfalls
- It is not a hard billing cap and does not automatically stop all resources.
- Budgets watch thresholds, Explorer analyzes spend, and Anomaly Detection finds deviations.
Key memory
Budgets for targets, Explorer for analysis, Anomaly Detection for unexpected deviation.