An enterprise has multiple business units operating under a consolidated billing family in AWS Organizations. Each team relies on specific cost allocation tags for their resources. The Solutions Architect must implement a cost management solution to meet two main requirements:
1. Receive alerts within 24 hours of any sudden, unexpected spending spikes that deviate from historical usage patterns.
2. Track monthly spend against a budget for each resource tag and receive email warnings when the forecasted cost is projected to exceed the allocated budget.
Which combination of AWS services or features will meet these requirements with the least administrative overhead? (Select TWO.)
- AWS Cost Anomaly Detection to monitor cost monitors and send alerts when abnormal spending patterns are detected.Answer
- AWS Budgets with cost filters based on cost allocation tags to trigger notifications when forecasted amounts exceed thresholds.Answer
- CAWS Cost Explorer with daily scheduled reports configured to email administrators when cost changes occur.
- DAWS Billing Conductor to define pro forma billing groups and set up billing alerts for each tag configuration.
- EAWS Budgets Reports configured to analyze historical anomalies and trigger warning notifications to email distribution lists.
Answer
AWS Cost Anomaly Detection should be used to detect sudden, unexpected spending spikes using machine learning, and AWS Budgets with cost filters should be used to track monthly tag-based spending against forecasted thresholds.
The correct options are the use of AWS Cost Anomaly Detection and AWS Budgets. AWS Cost Anomaly Detection satisfies the requirement to identify unexpected cost spikes that deviate from historical patterns within 24 hours using machine learning. AWS Budgets satisfies the requirement to monitor specific tag-based costs and alert proactively based on forecasted monthly spending before the budget threshold is actually crossed.
Step-by-Step Solution
Key Concept
Selecting the appropriate AWS cost management tool based on the required monitoring latency (proactive forecasting vs. dynamic ML anomaly detection vs. retrospective reporting).