What it is: Time-series performance data (CPUUtilization, RequestCount, Latency, etc.).

What it’s for:

• Observe system health over time.
• Trigger alarms and auto scaling based on thresholds.

Key ideas:

• Metrics can be AWS-provided or custom.
• Commonly used for scaling decisions (ASG target tracking, alarms).

Exam cues:

• “scale when CPU > 70%” → metrics + alarm/policy.

Hard words:

• *time-series* /ˈtaɪm ˌsɪriːz/: chuỗi theo thời gian
• *utilization* /ˌjuːtələˈzeɪʃn/: mức sử dụng
• *threshold* /ˈθreʃhoʊld/: ngưỡng