Basic Information
| Field | Value |
|---|---|
| English | Amazon CloudWatch |
| Full name | Amazon CloudWatch |
| Chinese | 监控、日志与可观测性服务 |
| Japanese | Amazon CloudWatch（モニタリング・ログ・可観測性サービス） |
| Exam frequency | ⭐⭐⭐⭐⭐ |
| Often confused with | AWS CloudTrail / AWS Config / AWS Health |
One-line summary
Collect metrics and logs, visualize dashboards, and trigger alarms to observe current system health.
Core capabilities and use cases
- Metrics quantify CPU, request count, latency, error rate, and custom measurements.
- Logs centralizes application, operating-system, Lambda, and service logs; Logs Insights queries them.
- Alarms evaluate metric conditions and can notify SNS, trigger Auto Scaling, or start automation.
- Dashboards visualize key information but do not create alerts by themselves.
- EC2 memory and other guest OS metrics commonly require the CloudWatch Agent.
Exam focus and common pitfalls
- CloudWatch answers how the system is performing now; CloudTrail answers who called an API; Config answers how configuration changed.
- Alarm states are OK, ALARM, and INSUFFICIENT_DATA.
- Metrics measure, alarms decide, SNS notifies, dashboards display, and logs support troubleshooting.
Key takeaway
Metrics measure, alarms decide, SNS notifies, dashboards display, and logs troubleshoot.