Performance metrics are quantitative measurements of system behavior, resource use, or performance over time. In AWS, services publish metrics to CloudWatch so you can inspect capacity, debug bottlenecks, define alarms, and trigger scaling actions. section 3.2.2 section 17.2.2
Apply metrics to practical operational questions:
- Inspect EC2 CPU, network, and disk usage to judge whether a virtual machine is nearing capacity; publish a custom memory metric when needed. section 3.2.2
- Monitor RDS storage, CPU, memory, disk queues, and related measurements, then create CloudWatch alarms for dangerous thresholds. section 10.7
- Use CloudWatch alarms with scaling policies to increase or decrease an Auto Scaling group’s EC2 capacity as workload changes. section 17.2.2
Scheduled website health check
A Lambda function sends an HTTP request and verifies expected response text. An EventBridge rule runs it every five minutes, while a CloudWatch alarm counts failures and sends email when the site is unavailable. section 6.2
