Amazon CloudWatch Metrics: Most AWS services provide free metrics

Enable detailed monitoring ($$$) if needed

EC2: CPUUtilization, NetworkIn, NetworkOut

(DEFAULT) EC2 instances collect metrics every 5 minutes.

You can increase it to every one minute ($$$)

CloudWatch does NOT have access to operating system metrics like memory consumption

Install CloudWatch agent to gather metrics around memory

ELB: HTTPCode_Target_2XX_Count, HTTPCode_Target_4XX_Count

DynamoDB: AccountProvisionedReadCapacityUtilization, ConsumedReadCapacityUnits

Lambda: Throttles, Errors, ConcurrentExecutions, Duration

API Gateway: 4XXError,5XXError, CacheHitCount, CacheMissCount, Count

IntegrationLatency (How long did the backend take to process?)

Latency (How long did the total client request to API Gateway take?)

Metrics exists only in the region in which they are created.